Table Of Contents

Search

Enter search terms or a module, class or function name.

What’s New

These are new features and improvements of note in each release.

v0.19.1 (November 3, 2016)

This is a minor bug-fix release from 0.19.0 and includes some small regression fixes, bug fixes and performance improvements. We recommend that all users upgrade to this version.

What’s new in v0.19.1

Performance Improvements

  • Fixed performance regression in factorization of Period data (GH14338)
  • Fixed performance regression in Series.asof(where) when where is a scalar (GH14461)
  • Improved performance in DataFrame.asof(where) when where is a scalar (GH14461)
  • Improved performance in .to_json() when lines=True (GH14408)
  • Improved performance in certain types of loc indexing with a MultiIndex (GH14551).

Bug Fixes

  • Source installs from PyPI will now again work without cython installed, as in previous versions (GH14204)
  • Compat with Cython 0.25 for building (GH14496)
  • Fixed regression where user-provided file handles were closed in read_csv (c engine) (GH14418).
  • Fixed regression in DataFrame.quantile when missing values where present in some columns (GH14357).
  • Fixed regression in Index.difference where the freq of a DatetimeIndex was incorrectly set (GH14323)
  • Added back pandas.core.common.array_equivalent with a deprecation warning (GH14555).
  • Bug in pd.read_csv for the C engine in which quotation marks were improperly parsed in skipped rows (GH14459)
  • Bug in pd.read_csv for Python 2.x in which Unicode quote characters were no longer being respected (GH14477)
  • Fixed regression in Index.append when categorical indices were appended (GH14545).
  • Fixed regression in pd.DataFrame where constructor fails when given dict with None value (GH14381)
  • Fixed regression in DatetimeIndex._maybe_cast_slice_bound when index is empty (GH14354).
  • Bug in localizing an ambiguous timezone when a boolean is passed (GH14402)
  • Bug in TimedeltaIndex addition with a Datetime-like object where addition overflow in the negative direction was not being caught (GH14068, GH14453)
  • Bug in string indexing against data with object Index may raise AttributeError (GH14424)
  • Corrrecly raise ValueError on empty input to pd.eval() and df.query() (GH13139)
  • Bug in RangeIndex.intersection when result is a empty set (GH14364).
  • Bug in groupby-transform broadcasting that could cause incorrect dtype coercion (GH14457)
  • Bug in Series.__setitem__ which allowed mutating read-only arrays (GH14359).
  • Bug in DataFrame.insert where multiple calls with duplicate columns can fail (GH14291)
  • pd.merge() will raise ValueError with non-boolean parameters in passed boolean type arguments (GH14434)
  • Bug in Timestamp where dates very near the minimum (1677-09) could underflow on creation (GH14415)
  • Bug in pd.concat where names of the keys were not propagated to the resulting MultiIndex (GH14252)
  • Bug in pd.concat where axis cannot take string parameters 'rows' or 'columns' (GH14369)
  • Bug in pd.concat with dataframes heterogeneous in length and tuple keys (GH14438)
  • Bug in MultiIndex.set_levels where illegal level values were still set after raising an error (GH13754)
  • Bug in DataFrame.to_json where lines=True and a value contained a } character (GH14391)
  • Bug in df.groupby causing an AttributeError when grouping a single index frame by a column and the index level (:issue`14327`)
  • Bug in df.groupby where TypeError raised when pd.Grouper(key=...) is passed in a list (GH14334)
  • Bug in pd.pivot_table may raise TypeError or ValueError when index or columns is not scalar and values is not specified (GH14380)

v0.19.0 (October 2, 2016)

This is a major release from 0.18.1 and includes number of API changes, several new features, enhancements, and performance improvements along with a large number of bug fixes. We recommend that all users upgrade to this version.

Highlights include:

  • merge_asof() for asof-style time-series joining, see here
  • .rolling() is now time-series aware, see here
  • read_csv() now supports parsing Categorical data, see here
  • A function union_categorical() has been added for combining categoricals, see here
  • PeriodIndex now has its own period dtype, and changed to be more consistent with other Index classes. See here
  • Sparse data structures gained enhanced support of int and bool dtypes, see here
  • Comparison operations with Series no longer ignores the index, see here for an overview of the API changes.
  • Introduction of a pandas development API for utility functions, see here.
  • Deprecation of Panel4D and PanelND. We recommend to represent these types of n-dimensional data with the xarray package.
  • Removal of the previously deprecated modules pandas.io.data, pandas.io.wb, pandas.tools.rplot.

Warning

pandas >= 0.19.0 will no longer silence numpy ufunc warnings upon import, see here.

New features

merge_asof for asof-style time-series joining

A long-time requested feature has been added through the merge_asof() function, to support asof style joining of time-series (GH1870, GH13695, GH13709, GH13902). Full documentation is here.

The merge_asof() performs an asof merge, which is similar to a left-join except that we match on nearest key rather than equal keys.

In [1]: left = pd.DataFrame({'a': [1, 5, 10],
   ...:                      'left_val': ['a', 'b', 'c']})
   ...: 

In [2]: right = pd.DataFrame({'a': [1, 2, 3, 6, 7],
   ...:                      'right_val': [1, 2, 3, 6, 7]})
   ...: 

In [3]: left
Out[3]: 
    a left_val
0   1        a
1   5        b
2  10        c

In [4]: right
Out[4]: 
   a  right_val
0  1          1
1  2          2
2  3          3
3  6          6
4  7          7

We typically want to match exactly when possible, and use the most recent value otherwise.

In [5]: pd.merge_asof(left, right, on='a')
Out[5]: 
    a left_val  right_val
0   1        a          1
1   5        b          3
2  10        c          7

We can also match rows ONLY with prior data, and not an exact match.

In [6]: pd.merge_asof(left, right, on='a', allow_exact_matches=False)
Out[6]: 
    a left_val  right_val
0   1        a        NaN
1   5        b        3.0
2  10        c        7.0

In a typical time-series example, we have trades and quotes and we want to asof-join them. This also illustrates using the by parameter to group data before merging.

In [7]: trades = pd.DataFrame({
   ...:     'time': pd.to_datetime(['20160525 13:30:00.023',
   ...:                             '20160525 13:30:00.038',
   ...:                             '20160525 13:30:00.048',
   ...:                             '20160525 13:30:00.048',
   ...:                             '20160525 13:30:00.048']),
   ...:     'ticker': ['MSFT', 'MSFT',
   ...:                'GOOG', 'GOOG', 'AAPL'],
   ...:     'price': [51.95, 51.95,
   ...:               720.77, 720.92, 98.00],
   ...:     'quantity': [75, 155,
   ...:                  100, 100, 100]},
   ...:     columns=['time', 'ticker', 'price', 'quantity'])
   ...: 

In [8]: quotes = pd.DataFrame({
   ...:     'time': pd.to_datetime(['20160525 13:30:00.023',
   ...:                             '20160525 13:30:00.023',
   ...:                             '20160525 13:30:00.030',
   ...:                             '20160525 13:30:00.041',
   ...:                             '20160525 13:30:00.048',
   ...:                             '20160525 13:30:00.049',
   ...:                             '20160525 13:30:00.072',
   ...:                             '20160525 13:30:00.075']),
   ...:     'ticker': ['GOOG', 'MSFT', 'MSFT',
   ...:                'MSFT', 'GOOG', 'AAPL', 'GOOG',
   ...:                'MSFT'],
   ...:     'bid': [720.50, 51.95, 51.97, 51.99,
   ...:             720.50, 97.99, 720.50, 52.01],
   ...:     'ask': [720.93, 51.96, 51.98, 52.00,
   ...:             720.93, 98.01, 720.88, 52.03]},
   ...:     columns=['time', 'ticker', 'bid', 'ask'])
   ...: 
In [9]: trades
Out[9]: 
                     time ticker   price  quantity
0 2016-05-25 13:30:00.023   MSFT   51.95        75
1 2016-05-25 13:30:00.038   MSFT   51.95       155
2 2016-05-25 13:30:00.048   GOOG  720.77       100
3 2016-05-25 13:30:00.048   GOOG  720.92       100
4 2016-05-25 13:30:00.048   AAPL   98.00       100

In [10]: quotes
Out[10]: 
                     time ticker     bid     ask
0 2016-05-25 13:30:00.023   GOOG  720.50  720.93
1 2016-05-25 13:30:00.023   MSFT   51.95   51.96
2 2016-05-25 13:30:00.030   MSFT   51.97   51.98
3 2016-05-25 13:30:00.041   MSFT   51.99   52.00
4 2016-05-25 13:30:00.048   GOOG  720.50  720.93
5 2016-05-25 13:30:00.049   AAPL   97.99   98.01
6 2016-05-25 13:30:00.072   GOOG  720.50  720.88
7 2016-05-25 13:30:00.075   MSFT   52.01   52.03

An asof merge joins on the on, typically a datetimelike field, which is ordered, and in this case we are using a grouper in the by field. This is like a left-outer join, except that forward filling happens automatically taking the most recent non-NaN value.

In [11]: pd.merge_asof(trades, quotes,
   ....:               on='time',
   ....:               by='ticker')
   ....: 
Out[11]: 
                     time ticker   price  quantity     bid     ask
0 2016-05-25 13:30:00.023   MSFT   51.95        75   51.95   51.96
1 2016-05-25 13:30:00.038   MSFT   51.95       155   51.97   51.98
2 2016-05-25 13:30:00.048   GOOG  720.77       100  720.50  720.93
3 2016-05-25 13:30:00.048   GOOG  720.92       100  720.50  720.93
4 2016-05-25 13:30:00.048   AAPL   98.00       100     NaN     NaN

This returns a merged DataFrame with the entries in the same order as the original left passed DataFrame (trades in this case), with the fields of the quotes merged.

.rolling() is now time-series aware

.rolling() objects are now time-series aware and can accept a time-series offset (or convertible) for the window argument (GH13327, GH12995). See the full documentation here.

In [12]: dft = pd.DataFrame({'B': [0, 1, 2, np.nan, 4]},
   ....:                    index=pd.date_range('20130101 09:00:00', periods=5, freq='s'))
   ....: 

In [13]: dft
Out[13]: 
                       B
2013-01-01 09:00:00  0.0
2013-01-01 09:00:01  1.0
2013-01-01 09:00:02  2.0
2013-01-01 09:00:03  NaN
2013-01-01 09:00:04  4.0

This is a regular frequency index. Using an integer window parameter works to roll along the window frequency.

In [14]: dft.rolling(2).sum()
Out[14]: 
                       B
2013-01-01 09:00:00  NaN
2013-01-01 09:00:01  1.0
2013-01-01 09:00:02  3.0
2013-01-01 09:00:03  NaN
2013-01-01 09:00:04  NaN

In [15]: dft.rolling(2, min_periods=1).sum()
Out[15]: 
                       B
2013-01-01 09:00:00  0.0
2013-01-01 09:00:01  1.0
2013-01-01 09:00:02  3.0
2013-01-01 09:00:03  2.0
2013-01-01 09:00:04  4.0

Specifying an offset allows a more intuitive specification of the rolling frequency.

In [16]: dft.rolling('2s').sum()
Out[16]: 
                       B
2013-01-01 09:00:00  0.0
2013-01-01 09:00:01  1.0
2013-01-01 09:00:02  3.0
2013-01-01 09:00:03  2.0
2013-01-01 09:00:04  4.0

Using a non-regular, but still monotonic index, rolling with an integer window does not impart any special calculation.

In [17]: dft = DataFrame({'B': [0, 1, 2, np.nan, 4]},
   ....:                 index = pd.Index([pd.Timestamp('20130101 09:00:00'),
   ....:                                   pd.Timestamp('20130101 09:00:02'),
   ....:                                   pd.Timestamp('20130101 09:00:03'),
   ....:                                   pd.Timestamp('20130101 09:00:05'),
   ....:                                   pd.Timestamp('20130101 09:00:06')],
   ....:                                  name='foo'))
   ....: 

In [18]: dft
Out[18]: 
                       B
foo                     
2013-01-01 09:00:00  0.0
2013-01-01 09:00:02  1.0
2013-01-01 09:00:03  2.0
2013-01-01 09:00:05  NaN
2013-01-01 09:00:06  4.0

In [19]: dft.rolling(2).sum()
Out[19]: 
                       B
foo                     
2013-01-01 09:00:00  NaN
2013-01-01 09:00:02  1.0
2013-01-01 09:00:03  3.0
2013-01-01 09:00:05  NaN
2013-01-01 09:00:06  NaN

Using the time-specification generates variable windows for this sparse data.

In [20]: dft.rolling('2s').sum()
Out[20]: 
                       B
foo                     
2013-01-01 09:00:00  0.0
2013-01-01 09:00:02  1.0
2013-01-01 09:00:03  3.0
2013-01-01 09:00:05  NaN
2013-01-01 09:00:06  4.0

Furthermore, we now allow an optional on parameter to specify a column (rather than the default of the index) in a DataFrame.

In [21]: dft = dft.reset_index()

In [22]: dft
Out[22]: 
                  foo    B
0 2013-01-01 09:00:00  0.0
1 2013-01-01 09:00:02  1.0
2 2013-01-01 09:00:03  2.0
3 2013-01-01 09:00:05  NaN
4 2013-01-01 09:00:06  4.0

In [23]: dft.rolling('2s', on='foo').sum()
Out[23]: 
                  foo    B
0 2013-01-01 09:00:00  0.0
1 2013-01-01 09:00:02  1.0
2 2013-01-01 09:00:03  3.0
3 2013-01-01 09:00:05  NaN
4 2013-01-01 09:00:06  4.0

read_csv has improved support for duplicate column names

Duplicate column names are now supported in read_csv() whether they are in the file or passed in as the names parameter (GH7160, GH9424)

In [24]: data = '0,1,2\n3,4,5'

In [25]: names = ['a', 'b', 'a']

Previous behavior:

In [2]: pd.read_csv(StringIO(data), names=names)
Out[2]:
   a  b  a
0  2  1  2
1  5  4  5

The first a column contained the same data as the second a column, when it should have contained the values [0, 3].

New behavior:

In [26]: pd.read_csv(StringIO(data), names=names)
Out[26]: 
   a  b  a.1
0  0  1    2
1  3  4    5

read_csv supports parsing Categorical directly

The read_csv() function now supports parsing a Categorical column when specified as a dtype (GH10153). Depending on the structure of the data, this can result in a faster parse time and lower memory usage compared to converting to Categorical after parsing. See the io docs here.

In [27]: data = 'col1,col2,col3\na,b,1\na,b,2\nc,d,3'

In [28]: pd.read_csv(StringIO(data))
Out[28]: 
  col1 col2  col3
0    a    b     1
1    a    b     2
2    c    d     3

In [29]: pd.read_csv(StringIO(data)).dtypes
Out[29]: 
col1    object
col2    object
col3     int64
dtype: object

In [30]: pd.read_csv(StringIO(data), dtype='category').dtypes
Out[30]: 
col1    category
col2    category
col3    category
dtype: object

Individual columns can be parsed as a Categorical using a dict specification

In [31]: pd.read_csv(StringIO(data), dtype={'col1': 'category'}).dtypes
Out[31]: 
col1    category
col2      object
col3       int64
dtype: object

Note

The resulting categories will always be parsed as strings (object dtype). If the categories are numeric they can be converted using the to_numeric() function, or as appropriate, another converter such as to_datetime().

In [32]: df = pd.read_csv(StringIO(data), dtype='category')

In [33]: df.dtypes
Out[33]: 
col1    category
col2    category
col3    category
dtype: object

In [34]: df['col3']
Out[34]: 
0    1
1    2
2    3
Name: col3, dtype: category
Categories (3, object): [1, 2, 3]

In [35]: df['col3'].cat.categories = pd.to_numeric(df['col3'].cat.categories)

In [36]: df['col3']
Out[36]: 
0    1
1    2
2    3
Name: col3, dtype: category
Categories (3, int64): [1, 2, 3]

Categorical Concatenation

  • A function union_categoricals() has been added for combining categoricals, see Unioning Categoricals (GH13361, GH:13763, issue:13846, GH14173)

    In [37]: from pandas.types.concat import union_categoricals
    
    In [38]: a = pd.Categorical(["b", "c"])
    
    In [39]: b = pd.Categorical(["a", "b"])
    
    In [40]: union_categoricals([a, b])
    Out[40]: 
    [b, c, a, b]
    Categories (3, object): [b, c, a]
    
  • concat and append now can concat category dtypes with different categories as object dtype (GH13524)

    In [41]: s1 = pd.Series(['a', 'b'], dtype='category')
    
    In [42]: s2 = pd.Series(['b', 'c'], dtype='category')
    

    Previous behavior:

    In [1]: pd.concat([s1, s2])
    ValueError: incompatible categories in categorical concat
    

    New behavior:

    In [43]: pd.concat([s1, s2])
    Out[43]: 
    0    a
    1    b
    0    b
    1    c
    dtype: object
    

Semi-Month Offsets

Pandas has gained new frequency offsets, SemiMonthEnd (‘SM’) and SemiMonthBegin (‘SMS’). These provide date offsets anchored (by default) to the 15th and end of month, and 15th and 1st of month respectively. (GH1543)

In [44]: from pandas.tseries.offsets import SemiMonthEnd, SemiMonthBegin

SemiMonthEnd:

In [45]: Timestamp('2016-01-01') + SemiMonthEnd()
Out[45]: Timestamp('2016-01-15 00:00:00')

In [46]: pd.date_range('2015-01-01', freq='SM', periods=4)
Out[46]: DatetimeIndex(['2015-01-15', '2015-01-31', '2015-02-15', '2015-02-28'], dtype='datetime64[ns]', freq='SM-15')

SemiMonthBegin:

In [47]: Timestamp('2016-01-01') + SemiMonthBegin()
Out[47]: Timestamp('2016-01-15 00:00:00')

In [48]: pd.date_range('2015-01-01', freq='SMS', periods=4)
Out[48]: DatetimeIndex(['2015-01-01', '2015-01-15', '2015-02-01', '2015-02-15'], dtype='datetime64[ns]', freq='SMS-15')

Using the anchoring suffix, you can also specify the day of month to use instead of the 15th.

In [49]: pd.date_range('2015-01-01', freq='SMS-16', periods=4)
Out[49]: DatetimeIndex(['2015-01-01', '2015-01-16', '2015-02-01', '2015-02-16'], dtype='datetime64[ns]', freq='SMS-16')

In [50]: pd.date_range('2015-01-01', freq='SM-14', periods=4)
Out[50]: DatetimeIndex(['2015-01-14', '2015-01-31', '2015-02-14', '2015-02-28'], dtype='datetime64[ns]', freq='SM-14')

New Index methods

The following methods and options are added to Index, to be more consistent with the Series and DataFrame API.

Index now supports the .where() function for same shape indexing (GH13170)

In [51]: idx = pd.Index(['a', 'b', 'c'])

In [52]: idx.where([True, False, True])
Out[52]: Index([u'a', nan, u'c'], dtype='object')

Index now supports .dropna() to exclude missing values (GH6194)

In [53]: idx = pd.Index([1, 2, np.nan, 4])

In [54]: idx.dropna()
Out[54]: Float64Index([1.0, 2.0, 4.0], dtype='float64')

For MultiIndex, values are dropped if any level is missing by default. Specifying how='all' only drops values where all levels are missing.

In [55]: midx = pd.MultiIndex.from_arrays([[1, 2, np.nan, 4],
   ....:                                     [1, 2, np.nan, np.nan]])
   ....: 

In [56]: midx
Out[56]: 
MultiIndex(levels=[[1, 2, 4], [1, 2]],
           labels=[[0, 1, -1, 2], [0, 1, -1, -1]])

In [57]: midx.dropna()
Out[57]: 
MultiIndex(levels=[[1, 2, 4], [1, 2]],
           labels=[[0, 1], [0, 1]])

In [58]: midx.dropna(how='all')
Out[58]: 
MultiIndex(levels=[[1, 2, 4], [1, 2]],
           labels=[[0, 1, 2], [0, 1, -1]])

Index now supports .str.extractall() which returns a DataFrame, see the docs here (GH10008, GH13156)

In [59]: idx = pd.Index(["a1a2", "b1", "c1"])

In [60]: idx.str.extractall("[ab](?P<digit>\d)")
Out[60]: 
        digit
  match      
0 0         1
  1         2
1 0         1

Index.astype() now accepts an optional boolean argument copy, which allows optional copying if the requirements on dtype are satisfied (GH13209)

Google BigQuery Enhancements

  • The read_gbq() method has gained the dialect argument to allow users to specify whether to use BigQuery’s legacy SQL or BigQuery’s standard SQL. See the docs for more details (GH13615).
  • The to_gbq() method now allows the DataFrame column order to differ from the destination table schema (GH11359).

Fine-grained numpy errstate

Previous versions of pandas would permanently silence numpy’s ufunc error handling when pandas was imported. Pandas did this in order to silence the warnings that would arise from using numpy ufuncs on missing data, which are usually represented as NaN s. Unfortunately, this silenced legitimate warnings arising in non-pandas code in the application. Starting with 0.19.0, pandas will use the numpy.errstate context manager to silence these warnings in a more fine-grained manner, only around where these operations are actually used in the pandas codebase. (GH13109, GH13145)

After upgrading pandas, you may see new RuntimeWarnings being issued from your code. These are likely legitimate, and the underlying cause likely existed in the code when using previous versions of pandas that simply silenced the warning. Use numpy.errstate around the source of the RuntimeWarning to control how these conditions are handled.

get_dummies now returns integer dtypes

The pd.get_dummies function now returns dummy-encoded columns as small integers, rather than floats (GH8725). This should provide an improved memory footprint.

Previous behavior:

In [1]: pd.get_dummies(['a', 'b', 'a', 'c']).dtypes

Out[1]:
a    float64
b    float64
c    float64
dtype: object

New behavior:

In [61]: pd.get_dummies(['a', 'b', 'a', 'c']).dtypes
Out[61]: 
a    uint8
b    uint8
c    uint8
dtype: object

Downcast values to smallest possible dtype in to_numeric

pd.to_numeric() now accepts a downcast parameter, which will downcast the data if possible to smallest specified numerical dtype (GH13352)

In [62]: s = ['1', 2, 3]

In [63]: pd.to_numeric(s, downcast='unsigned')
Out[63]: array([1, 2, 3], dtype=uint8)

In [64]: pd.to_numeric(s, downcast='integer')
Out[64]: array([1, 2, 3], dtype=int8)

pandas development API

As part of making pandas API more uniform and accessible in the future, we have created a standard sub-package of pandas, pandas.api to hold public API’s. We are starting by exposing type introspection functions in pandas.api.types. More sub-packages and officially sanctioned API’s will be published in future versions of pandas (GH13147, GH13634)

The following are now part of this API:

In [65]: import pprint

In [66]: from pandas.api import types

In [67]: funcs = [ f for f in dir(types) if not f.startswith('_') ]

In [68]: pprint.pprint(funcs)
['is_any_int_dtype',
 'is_bool',
 'is_bool_dtype',
 'is_categorical',
 'is_categorical_dtype',
 'is_complex',
 'is_complex_dtype',
 'is_datetime64_any_dtype',
 'is_datetime64_dtype',
 'is_datetime64_ns_dtype',
 'is_datetime64tz_dtype',
 'is_datetimetz',
 'is_dict_like',
 'is_dtype_equal',
 'is_extension_type',
 'is_float',
 'is_float_dtype',
 'is_floating_dtype',
 'is_hashable',
 'is_int64_dtype',
 'is_integer',
 'is_integer_dtype',
 'is_iterator',
 'is_list_like',
 'is_named_tuple',
 'is_number',
 'is_numeric_dtype',
 'is_object_dtype',
 'is_period',
 'is_period_dtype',
 'is_re',
 'is_re_compilable',
 'is_scalar',
 'is_sequence',
 'is_sparse',
 'is_string_dtype',
 'is_timedelta64_dtype',
 'is_timedelta64_ns_dtype',
 'pandas_dtype']

Note

Calling these functions from the internal module pandas.core.common will now show a DeprecationWarning (GH13990)

Other enhancements

  • Timestamp can now accept positional and keyword parameters similar to datetime.datetime() (GH10758, GH11630)

    In [69]: pd.Timestamp(2012, 1, 1)
    Out[69]: Timestamp('2012-01-01 00:00:00')
    
    In [70]: pd.Timestamp(year=2012, month=1, day=1, hour=8, minute=30)
    Out[70]: Timestamp('2012-01-01 08:30:00')
    
  • The .resample() function now accepts a on= or level= parameter for resampling on a datetimelike column or MultiIndex level (GH13500)

    In [71]: df = pd.DataFrame({'date': pd.date_range('2015-01-01', freq='W', periods=5),
       ....:                    'a': np.arange(5)},
       ....:                   index=pd.MultiIndex.from_arrays([
       ....:                            [1,2,3,4,5],
       ....:                            pd.date_range('2015-01-01', freq='W', periods=5)],
       ....:                        names=['v','d']))
       ....: 
    
    In [72]: df
    Out[72]: 
                  a       date
    v d                       
    1 2015-01-04  0 2015-01-04
    2 2015-01-11  1 2015-01-11
    3 2015-01-18  2 2015-01-18
    4 2015-01-25  3 2015-01-25
    5 2015-02-01  4 2015-02-01
    
    In [73]: df.resample('M', on='date').sum()
    Out[73]: 
                a
    date         
    2015-01-31  6
    2015-02-28  4
    
    In [74]: df.resample('M', level='d').sum()
    Out[74]: 
                a
    d            
    2015-01-31  6
    2015-02-28  4
    
  • The .get_credentials() method of GbqConnector can now first try to fetch the application default credentials. See the docs for more details (GH13577).

  • The .tz_localize() method of DatetimeIndex and Timestamp has gained the errors keyword, so you can potentially coerce nonexistent timestamps to NaT. The default behavior remains to raising a NonExistentTimeError (GH13057)

  • .to_hdf/read_hdf() now accept path objects (e.g. pathlib.Path, py.path.local) for the file path (GH11773)

  • The pd.read_csv() with engine='python' has gained support for the decimal (GH12933), na_filter (GH13321) and the memory_map option (GH13381).

  • Consistent with the Python API, pd.read_csv() will now interpret +inf as positive infinity (GH13274)

  • The pd.read_html() has gained support for the na_values, converters, keep_default_na options (GH13461)

  • Categorical.astype() now accepts an optional boolean argument copy, effective when dtype is categorical (GH13209)

  • DataFrame has gained the .asof() method to return the last non-NaN values according to the selected subset (GH13358)

  • The DataFrame constructor will now respect key ordering if a list of OrderedDict objects are passed in (GH13304)

  • pd.read_html() has gained support for the decimal option (GH12907)

  • Series has gained the properties .is_monotonic, .is_monotonic_increasing, .is_monotonic_decreasing, similar to Index (GH13336)

  • DataFrame.to_sql() now allows a single value as the SQL type for all columns (GH11886).

  • Series.append now supports the ignore_index option (GH13677)

  • .to_stata() and StataWriter can now write variable labels to Stata dta files using a dictionary to make column names to labels (GH13535, GH13536)

  • .to_stata() and StataWriter will automatically convert datetime64[ns] columns to Stata format %tc, rather than raising a ValueError (GH12259)

  • read_stata() and StataReader raise with a more explicit error message when reading Stata files with repeated value labels when convert_categoricals=True (GH13923)

  • DataFrame.style will now render sparsified MultiIndexes (GH11655)

  • DataFrame.style will now show column level names (e.g. DataFrame.columns.names) (GH13775)

  • DataFrame has gained support to re-order the columns based on the values in a row using df.sort_values(by='...', axis=1) (GH10806)

    In [75]: df = pd.DataFrame({'A': [2, 7], 'B': [3, 5], 'C': [4, 8]},
       ....:                   index=['row1', 'row2'])
       ....: 
    
    In [76]: df
    Out[76]: 
          A  B  C
    row1  2  3  4
    row2  7  5  8
    
    In [77]: df.sort_values(by='row2', axis=1)
    Out[77]: 
          B  A  C
    row1  3  2  4
    row2  5  7  8
    
  • Added documentation to I/O regarding the perils of reading in columns with mixed dtypes and how to handle it (GH13746)

  • to_html() now has a border argument to control the value in the opening <table> tag. The default is the value of the html.border option, which defaults to 1. This also affects the notebook HTML repr, but since Jupyter’s CSS includes a border-width attribute, the visual effect is the same. (GH11563).

  • Raise ImportError in the sql functions when sqlalchemy is not installed and a connection string is used (GH11920).

  • Compatibility with matplotlib 2.0. Older versions of pandas should also work with matplotlib 2.0 (GH13333)

  • Timestamp, Period, DatetimeIndex, PeriodIndex and .dt accessor have gained a .is_leap_year property to check whether the date belongs to a leap year. (GH13727)

  • astype() will now accept a dict of column name to data types mapping as the dtype argument. (GH12086)

  • The pd.read_json and DataFrame.to_json has gained support for reading and writing json lines with lines option see Line delimited json (GH9180)

  • read_excel() now supports the true_values and false_values keyword arguments (GH13347)

  • groupby() will now accept a scalar and a single-element list for specifying level on a non-MultiIndex grouper. (GH13907)

  • Non-convertible dates in an excel date column will be returned without conversion and the column will be object dtype, rather than raising an exception (GH10001).

  • pd.Timedelta(None) is now accepted and will return NaT, mirroring pd.Timestamp (GH13687)

  • pd.read_stata() can now handle some format 111 files, which are produced by SAS when generating Stata dta files (GH11526)

  • Series and Index now support divmod which will return a tuple of series or indices. This behaves like a standard binary operator with regards to broadcasting rules (GH14208).

API changes

Series.tolist() will now return Python types

Series.tolist() will now return Python types in the output, mimicking NumPy .tolist() behavior (GH10904)

In [78]: s = pd.Series([1,2,3])

Previous behavior:

In [7]: type(s.tolist()[0])
Out[7]:
 <class 'numpy.int64'>

New behavior:

In [79]: type(s.tolist()[0])
Out[79]: int

Series operators for different indexes

Following Series operators have been changed to make all operators consistent, including DataFrame (GH1134, GH4581, GH13538)

  • Series comparison operators now raise ValueError when index are different.
  • Series logical operators align both index of left and right hand side.

Warning

Until 0.18.1, comparing Series with the same length, would succeed even if the .index are different (the result ignores .index). As of 0.19.0, this will raises ValueError to be more strict. This section also describes how to keep previous behavior or align different indexes, using the flexible comparison methods like .eq.

As a result, Series and DataFrame operators behave as below:

Arithmetic operators

Arithmetic operators align both index (no changes).

In [80]: s1 = pd.Series([1, 2, 3], index=list('ABC'))

In [81]: s2 = pd.Series([2, 2, 2], index=list('ABD'))

In [82]: s1 + s2
Out[82]: 
A    3.0
B    4.0
C    NaN
D    NaN
dtype: float64

In [83]: df1 = pd.DataFrame([1, 2, 3], index=list('ABC'))

In [84]: df2 = pd.DataFrame([2, 2, 2], index=list('ABD'))

In [85]: df1 + df2
Out[85]: 
     0
A  3.0
B  4.0
C  NaN
D  NaN
Comparison operators

Comparison operators raise ValueError when .index are different.

Previous Behavior (Series):

Series compared values ignoring the .index as long as both had the same length:

In [1]: s1 == s2
Out[1]:
A    False
B     True
C    False
dtype: bool

New behavior (Series):

In [2]: s1 == s2
Out[2]:
ValueError: Can only compare identically-labeled Series objects

Note

To achieve the same result as previous versions (compare values based on locations ignoring .index), compare both .values.

In [86]: s1.values == s2.values
Out[86]: array([False,  True, False], dtype=bool)

If you want to compare Series aligning its .index, see flexible comparison methods section below:

In [87]: s1.eq(s2)
Out[87]: 
A    False
B     True
C    False
D    False
dtype: bool

Current Behavior (DataFrame, no change):

In [3]: df1 == df2
Out[3]:
ValueError: Can only compare identically-labeled DataFrame objects
Logical operators

Logical operators align both .index of left and right hand side.

Previous behavior (Series), only left hand side index was kept:

In [4]: s1 = pd.Series([True, False, True], index=list('ABC'))
In [5]: s2 = pd.Series([True, True, True], index=list('ABD'))
In [6]: s1 & s2
Out[6]:
A     True
B    False
C    False
dtype: bool

New behavior (Series):

In [88]: s1 = pd.Series([True, False, True], index=list('ABC'))

In [89]: s2 = pd.Series([True, True, True], index=list('ABD'))

In [90]: s1 & s2
Out[90]: 
A     True
B    False
C    False
D    False
dtype: bool

Note

Series logical operators fill a NaN result with False.

Note

To achieve the same result as previous versions (compare values based on only left hand side index), you can use reindex_like:

In [91]: s1 & s2.reindex_like(s1)
Out[91]: 
A     True
B    False
C    False
dtype: bool

Current Behavior (DataFrame, no change):

In [92]: df1 = pd.DataFrame([True, False, True], index=list('ABC'))

In [93]: df2 = pd.DataFrame([True, True, True], index=list('ABD'))

In [94]: df1 & df2
Out[94]: 
       0
A   True
B  False
C    NaN
D    NaN
Flexible comparison methods

Series flexible comparison methods like eq, ne, le, lt, ge and gt now align both index. Use these operators if you want to compare two Series which has the different index.

In [95]: s1 = pd.Series([1, 2, 3], index=['a', 'b', 'c'])

In [96]: s2 = pd.Series([2, 2, 2], index=['b', 'c', 'd'])

In [97]: s1.eq(s2)
Out[97]: 
a    False
b     True
c    False
d    False
dtype: bool

In [98]: s1.ge(s2)
Out[98]: 
a    False
b     True
c     True
d    False
dtype: bool

Previously, this worked the same as comparison operators (see above).

Series type promotion on assignment

A Series will now correctly promote its dtype for assignment with incompat values to the current dtype (GH13234)

In [99]: s = pd.Series()

Previous behavior:

In [2]: s["a"] = pd.Timestamp("2016-01-01")

In [3]: s["b"] = 3.0
TypeError: invalid type promotion

New behavior:

In [100]: s["a"] = pd.Timestamp("2016-01-01")

In [101]: s["b"] = 3.0

In [102]: s
Out[102]: 
a    2016-01-01 00:00:00
b                      3
dtype: object

In [103]: s.dtype
Out[103]: dtype('O')

.to_datetime() changes

Previously if .to_datetime() encountered mixed integers/floats and strings, but no datetimes with errors='coerce' it would convert all to NaT.

Previous behavior:

In [2]: pd.to_datetime([1, 'foo'], errors='coerce')
Out[2]: DatetimeIndex(['NaT', 'NaT'], dtype='datetime64[ns]', freq=None)

Current behavior:

This will now convert integers/floats with the default unit of ns.

In [104]: pd.to_datetime([1, 'foo'], errors='coerce')
Out[104]: DatetimeIndex(['1970-01-01 00:00:00.000000001', 'NaT'], dtype='datetime64[ns]', freq=None)

Bug fixes related to .to_datetime():

  • Bug in pd.to_datetime() when passing integers or floats, and no unit and errors='coerce' (GH13180).
  • Bug in pd.to_datetime() when passing invalid datatypes (e.g. bool); will now respect the errors keyword (GH13176)
  • Bug in pd.to_datetime() which overflowed on int8, and int16 dtypes (GH13451)
  • Bug in pd.to_datetime() raise AttributeError with NaN and the other string is not valid when errors='ignore' (GH12424)
  • Bug in pd.to_datetime() did not cast floats correctly when unit was specified, resulting in truncated datetime (GH13834)

Merging changes

Merging will now preserve the dtype of the join keys (GH8596)

In [105]: df1 = pd.DataFrame({'key': [1], 'v1': [10]})

In [106]: df1
Out[106]: 
   key  v1
0    1  10

In [107]: df2 = pd.DataFrame({'key': [1, 2], 'v1': [20, 30]})

In [108]: df2
Out[108]: 
   key  v1
0    1  20
1    2  30

Previous behavior:

In [5]: pd.merge(df1, df2, how='outer')
Out[5]:
   key    v1
0  1.0  10.0
1  1.0  20.0
2  2.0  30.0

In [6]: pd.merge(df1, df2, how='outer').dtypes
Out[6]:
key    float64
v1     float64
dtype: object

New behavior:

We are able to preserve the join keys

In [109]: pd.merge(df1, df2, how='outer')
Out[109]: 
   key  v1
0    1  10
1    1  20
2    2  30

In [110]: pd.merge(df1, df2, how='outer').dtypes
Out[110]: 
key    int64
v1     int64
dtype: object

Of course if you have missing values that are introduced, then the resulting dtype will be upcast, which is unchanged from previous.

In [111]: pd.merge(df1, df2, how='outer', on='key')
Out[111]: 
   key  v1_x  v1_y
0    1  10.0    20
1    2   NaN    30

In [112]: pd.merge(df1, df2, how='outer', on='key').dtypes
Out[112]: 
key       int64
v1_x    float64
v1_y      int64
dtype: object

.describe() changes

Percentile identifiers in the index of a .describe() output will now be rounded to the least precision that keeps them distinct (GH13104)

In [113]: s = pd.Series([0, 1, 2, 3, 4])

In [114]: df = pd.DataFrame([0, 1, 2, 3, 4])

Previous behavior:

The percentiles were rounded to at most one decimal place, which could raise ValueError for a data frame if the percentiles were duplicated.

In [3]: s.describe(percentiles=[0.0001, 0.0005, 0.001, 0.999, 0.9995, 0.9999])
Out[3]:
count     5.000000
mean      2.000000
std       1.581139
min       0.000000
0.0%      0.000400
0.1%      0.002000
0.1%      0.004000
50%       2.000000
99.9%     3.996000
100.0%    3.998000
100.0%    3.999600
max       4.000000
dtype: float64

In [4]: df.describe(percentiles=[0.0001, 0.0005, 0.001, 0.999, 0.9995, 0.9999])
Out[4]:
...
ValueError: cannot reindex from a duplicate axis

New behavior:

In [115]: s.describe(percentiles=[0.0001, 0.0005, 0.001, 0.999, 0.9995, 0.9999])
Out[115]: 
count     5.000000
mean      2.000000
std       1.581139
min       0.000000
0.01%     0.000400
0.05%     0.002000
0.1%      0.004000
50%       2.000000
99.9%     3.996000
99.95%    3.998000
99.99%    3.999600
max       4.000000
dtype: float64

In [116]: df.describe(percentiles=[0.0001, 0.0005, 0.001, 0.999, 0.9995, 0.9999])
Out[116]: 
               0
count   5.000000
mean    2.000000
std     1.581139
min     0.000000
0.01%   0.000400
0.05%   0.002000
0.1%    0.004000
50%     2.000000
99.9%   3.996000
99.95%  3.998000
99.99%  3.999600
max     4.000000

Furthermore:

  • Passing duplicated percentiles will now raise a ValueError.
  • Bug in .describe() on a DataFrame with a mixed-dtype column index, which would previously raise a TypeError (GH13288)

Period changes

PeriodIndex now has period dtype

PeriodIndex now has its own period dtype. The period dtype is a pandas extension dtype like category or the timezone aware dtype (datetime64[ns, tz]) (GH13941). As a consequence of this change, PeriodIndex no longer has an integer dtype:

Previous behavior:

In [1]: pi = pd.PeriodIndex(['2016-08-01'], freq='D')

In [2]: pi
Out[2]: PeriodIndex(['2016-08-01'], dtype='int64', freq='D')

In [3]: pd.api.types.is_integer_dtype(pi)
Out[3]: True

In [4]: pi.dtype
Out[4]: dtype('int64')

New behavior:

In [117]: pi = pd.PeriodIndex(['2016-08-01'], freq='D')

In [118]: pi
Out[118]: PeriodIndex(['2016-08-01'], dtype='period[D]', freq='D')

In [119]: pd.api.types.is_integer_dtype(pi)
Out[119]: False

In [120]: pd.api.types.is_period_dtype(pi)
Out[120]: True

In [121]: pi.dtype
Out[121]: period[D]

In [122]: type(pi.dtype)
Out[122]: pandas.types.dtypes.PeriodDtype
Period('NaT') now returns pd.NaT

Previously, Period has its own Period('NaT') representation different from pd.NaT. Now Period('NaT') has been changed to return pd.NaT. (GH12759, GH13582)

Previous behavior:

In [5]: pd.Period('NaT', freq='D')
Out[5]: Period('NaT', 'D')

New behavior:

These result in pd.NaT without providing freq option.

In [123]: pd.Period('NaT')
Out[123]: NaT

In [124]: pd.Period(None)
Out[124]: NaT

To be compatible with Period addition and subtraction, pd.NaT now supports addition and subtraction with int. Previously it raised ValueError.

Previous behavior:

In [5]: pd.NaT + 1
...
ValueError: Cannot add integral value to Timestamp without freq.

New behavior:

In [125]: pd.NaT + 1
Out[125]: NaT

In [126]: pd.NaT - 1
Out[126]: NaT
PeriodIndex.values now returns array of Period object

.values is changed to return an array of Period objects, rather than an array of integers (GH13988).

Previous behavior:

In [6]: pi = pd.PeriodIndex(['2011-01', '2011-02'], freq='M')
In [7]: pi.values
array([492, 493])

New behavior:

In [127]: pi = pd.PeriodIndex(['2011-01', '2011-02'], freq='M')

In [128]: pi.values
Out[128]: array([Period('2011-01', 'M'), Period('2011-02', 'M')], dtype=object)

Index + / - no longer used for set operations

Addition and subtraction of the base Index type and of DatetimeIndex (not the numeric index types) previously performed set operations (set union and difference). This behavior was already deprecated since 0.15.0 (in favor using the specific .union() and .difference() methods), and is now disabled. When possible, + and - are now used for element-wise operations, for example for concatenating strings or subtracting datetimes (GH8227, GH14127).

Previous behavior:

In [1]: pd.Index(['a', 'b']) + pd.Index(['a', 'c'])
FutureWarning: using '+' to provide set union with Indexes is deprecated, use '|' or .union()
Out[1]: Index(['a', 'b', 'c'], dtype='object')

New behavior: the same operation will now perform element-wise addition:

In [129]: pd.Index(['a', 'b']) + pd.Index(['a', 'c'])
Out[129]: Index([u'aa', u'bc'], dtype='object')

Note that numeric Index objects already performed element-wise operations. For example, the behavior of adding two integer Indexes is unchanged. The base Index is now made consistent with this behavior.

In [130]: pd.Index([1, 2, 3]) + pd.Index([2, 3, 4])
Out[130]: Int64Index([3, 5, 7], dtype='int64')

Further, because of this change, it is now possible to subtract two DatetimeIndex objects resulting in a TimedeltaIndex:

Previous behavior:

In [1]: pd.DatetimeIndex(['2016-01-01', '2016-01-02']) - pd.DatetimeIndex(['2016-01-02', '2016-01-03'])
FutureWarning: using '-' to provide set differences with datetimelike Indexes is deprecated, use .difference()
Out[1]: DatetimeIndex(['2016-01-01'], dtype='datetime64[ns]', freq=None)

New behavior:

In [131]: pd.DatetimeIndex(['2016-01-01', '2016-01-02']) - pd.DatetimeIndex(['2016-01-02', '2016-01-03'])
Out[131]: TimedeltaIndex(['-1 days', '-1 days'], dtype='timedelta64[ns]', freq=None)

Index.difference and .symmetric_difference changes

Index.difference and Index.symmetric_difference will now, more consistently, treat NaN values as any other values. (GH13514)

In [132]: idx1 = pd.Index([1, 2, 3, np.nan])

In [133]: idx2 = pd.Index([0, 1, np.nan])

Previous behavior:

In [3]: idx1.difference(idx2)
Out[3]: Float64Index([nan, 2.0, 3.0], dtype='float64')

In [4]: idx1.symmetric_difference(idx2)
Out[4]: Float64Index([0.0, nan, 2.0, 3.0], dtype='float64')

New behavior:

In [134]: idx1.difference(idx2)
Out[134]: Float64Index([2.0, 3.0], dtype='float64')

In [135]: idx1.symmetric_difference(idx2)
Out[135]: Float64Index([0.0, 2.0, 3.0], dtype='float64')

Index.unique consistently returns Index

Index.unique() now returns unique values as an Index of the appropriate dtype. (GH13395). Previously, most Index classes returned np.ndarray, and DatetimeIndex, TimedeltaIndex and PeriodIndex returned Index to keep metadata like timezone.

Previous behavior:

In [1]: pd.Index([1, 2, 3]).unique()
Out[1]: array([1, 2, 3])

In [2]: pd.DatetimeIndex(['2011-01-01', '2011-01-02', '2011-01-03'], tz='Asia/Tokyo').unique()
Out[2]:
DatetimeIndex(['2011-01-01 00:00:00+09:00', '2011-01-02 00:00:00+09:00',
               '2011-01-03 00:00:00+09:00'],
              dtype='datetime64[ns, Asia/Tokyo]', freq=None)

New behavior:

In [136]: pd.Index([1, 2, 3]).unique()
Out[136]: Int64Index([1, 2, 3], dtype='int64')

In [137]: pd.DatetimeIndex(['2011-01-01', '2011-01-02', '2011-01-03'], tz='Asia/Tokyo').unique()
Out[137]: 
DatetimeIndex(['2011-01-01 00:00:00+09:00', '2011-01-02 00:00:00+09:00',
               '2011-01-03 00:00:00+09:00'],
              dtype='datetime64[ns, Asia/Tokyo]', freq=None)

MultiIndex constructors, groupby and set_index preserve categorical dtypes

MultiIndex.from_arrays and MultiIndex.from_product will now preserve categorical dtype in MultiIndex levels (GH13743, GH13854).

In [138]: cat = pd.Categorical(['a', 'b'], categories=list("bac"))

In [139]: lvl1 = ['foo', 'bar']

In [140]: midx = pd.MultiIndex.from_arrays([cat, lvl1])

In [141]: midx
Out[141]: 
MultiIndex(levels=[[u'b', u'a', u'c'], [u'bar', u'foo']],
           labels=[[1, 0], [1, 0]])

Previous behavior:

In [4]: midx.levels[0]
Out[4]: Index(['b', 'a', 'c'], dtype='object')

In [5]: midx.get_level_values[0]
Out[5]: Index(['a', 'b'], dtype='object')

New behavior: the single level is now a CategoricalIndex:

In [142]: midx.levels[0]
Out[142]: CategoricalIndex([u'b', u'a', u'c'], categories=[u'b', u'a', u'c'], ordered=False, dtype='category')

In [143]: midx.get_level_values(0)
Out[143]: CategoricalIndex([u'a', u'b'], categories=[u'b', u'a', u'c'], ordered=False, dtype='category')

An analogous change has been made to MultiIndex.from_product. As a consequence, groupby and set_index also preserve categorical dtypes in indexes

In [144]: df = pd.DataFrame({'A': [0, 1], 'B': [10, 11], 'C': cat})

In [145]: df_grouped = df.groupby(by=['A', 'C']).first()

In [146]: df_set_idx = df.set_index(['A', 'C'])

Previous behavior:

In [11]: df_grouped.index.levels[1]
Out[11]: Index(['b', 'a', 'c'], dtype='object', name='C')
In [12]: df_grouped.reset_index().dtypes
Out[12]:
A      int64
C     object
B    float64
dtype: object

In [13]: df_set_idx.index.levels[1]
Out[13]: Index(['b', 'a', 'c'], dtype='object', name='C')
In [14]: df_set_idx.reset_index().dtypes
Out[14]:
A      int64
C     object
B      int64
dtype: object

New behavior:

In [147]: df_grouped.index.levels[1]
Out[147]: CategoricalIndex([u'b', u'a', u'c'], categories=[u'b', u'a', u'c'], ordered=False, name=u'C', dtype='category')

In [148]: df_grouped.reset_index().dtypes
Out[148]: 
A       int64
C    category
B     float64
dtype: object

In [149]: df_set_idx.index.levels[1]
Out[149]: CategoricalIndex([u'b', u'a', u'c'], categories=[u'b', u'a', u'c'], ordered=False, name=u'C', dtype='category')

In [150]: df_set_idx.reset_index().dtypes
Out[150]: 
A       int64
C    category
B       int64
dtype: object

read_csv will progressively enumerate chunks

When read_csv() is called with chunksize=n and without specifying an index, each chunk used to have an independently generated index from 0 to n-1. They are now given instead a progressive index, starting from 0 for the first chunk, from n for the second, and so on, so that, when concatenated, they are identical to the result of calling read_csv() without the chunksize= argument (GH12185).

In [151]: data = 'A,B\n0,1\n2,3\n4,5\n6,7'

Previous behavior:

In [2]: pd.concat(pd.read_csv(StringIO(data), chunksize=2))
Out[2]:
   A  B
0  0  1
1  2  3
0  4  5
1  6  7

New behavior:

In [152]: pd.concat(pd.read_csv(StringIO(data), chunksize=2))
Out[152]: 
   A  B
0  0  1
1  2  3
2  4  5
3  6  7

Sparse Changes

These changes allow pandas to handle sparse data with more dtypes, and for work to make a smoother experience with data handling.

int64 and bool support enhancements

Sparse data structures now gained enhanced support of int64 and bool dtype (GH667, GH13849).

Previously, sparse data were float64 dtype by default, even if all inputs were of int or bool dtype. You had to specify dtype explicitly to create sparse data with int64 dtype. Also, fill_value had to be specified explicitly because the default was np.nan which doesn’t appear in int64 or bool data.

In [1]: pd.SparseArray([1, 2, 0, 0])
Out[1]:
[1.0, 2.0, 0.0, 0.0]
Fill: nan
IntIndex
Indices: array([0, 1, 2, 3], dtype=int32)

# specifying int64 dtype, but all values are stored in sp_values because
# fill_value default is np.nan
In [2]: pd.SparseArray([1, 2, 0, 0], dtype=np.int64)
Out[2]:
[1, 2, 0, 0]
Fill: nan
IntIndex
Indices: array([0, 1, 2, 3], dtype=int32)

In [3]: pd.SparseArray([1, 2, 0, 0], dtype=np.int64, fill_value=0)
Out[3]:
[1, 2, 0, 0]
Fill: 0
IntIndex
Indices: array([0, 1], dtype=int32)

As of v0.19.0, sparse data keeps the input dtype, and uses more appropriate fill_value defaults (0 for int64 dtype, False for bool dtype).

In [153]: pd.SparseArray([1, 2, 0, 0], dtype=np.int64)
Out[153]: 
[1, 2, 0, 0]
Fill: 0
IntIndex
Indices: array([0, 1], dtype=int32)

In [154]: pd.SparseArray([True, False, False, False])
Out[154]: 
[True, False, False, False]
Fill: False
IntIndex
Indices: array([0], dtype=int32)

See the docs for more details.

Operators now preserve dtypes
  • Sparse data structure now can preserve dtype after arithmetic ops (GH13848)

    In [155]: s = pd.SparseSeries([0, 2, 0, 1], fill_value=0, dtype=np.int64)
    
    In [156]: s.dtype
    Out[156]: dtype('int64')
    
    In [157]: s + 1
    Out[157]: 
    0    1
    1    3
    2    1
    3    2
    dtype: int64
    BlockIndex
    Block locations: array([1, 3], dtype=int32)
    Block lengths: array([1, 1], dtype=int32)
    
  • Sparse data structure now support astype to convert internal dtype (GH13900)

    In [158]: s = pd.SparseSeries([1., 0., 2., 0.], fill_value=0)
    
    In [159]: s
    Out[159]: 
    0    1.0
    1    0.0
    2    2.0
    3    0.0
    dtype: float64
    BlockIndex
    Block locations: array([0, 2], dtype=int32)
    Block lengths: array([1, 1], dtype=int32)
    
    In [160]: s.astype(np.int64)
    Out[160]: 
    0    1
    1    0
    2    2
    3    0
    dtype: int64
    BlockIndex
    Block locations: array([0, 2], dtype=int32)
    Block lengths: array([1, 1], dtype=int32)
    

    astype fails if data contains values which cannot be converted to specified dtype. Note that the limitation is applied to fill_value which default is np.nan.

    In [7]: pd.SparseSeries([1., np.nan, 2., np.nan], fill_value=np.nan).astype(np.int64)
    Out[7]:
    ValueError: unable to coerce current fill_value nan to int64 dtype
    
Other sparse fixes
  • Subclassed SparseDataFrame and SparseSeries now preserve class types when slicing or transposing. (GH13787)
  • SparseArray with bool dtype now supports logical (bool) operators (GH14000)
  • Bug in SparseSeries with MultiIndex [] indexing may raise IndexError (GH13144)
  • Bug in SparseSeries with MultiIndex [] indexing result may have normal Index (GH13144)
  • Bug in SparseDataFrame in which axis=None did not default to axis=0 (GH13048)
  • Bug in SparseSeries and SparseDataFrame creation with object dtype may raise TypeError (GH11633)
  • Bug in SparseDataFrame doesn’t respect passed SparseArray or SparseSeries ‘s dtype and fill_value (GH13866)
  • Bug in SparseArray and SparseSeries don’t apply ufunc to fill_value (GH13853)
  • Bug in SparseSeries.abs incorrectly keeps negative fill_value (GH13853)
  • Bug in single row slicing on multi-type SparseDataFrame s, types were previously forced to float (GH13917)
  • Bug in SparseSeries slicing changes integer dtype to float (GH8292)
  • Bug in SparseDataFarme comparison ops may raise TypeError (GH13001)
  • Bug in SparseDataFarme.isnull raises ValueError (GH8276)
  • Bug in SparseSeries representation with bool dtype may raise IndexError (GH13110)
  • Bug in SparseSeries and SparseDataFrame of bool or int64 dtype may display its values like float64 dtype (GH13110)
  • Bug in sparse indexing using SparseArray with bool dtype may return incorrect result (GH13985)
  • Bug in SparseArray created from SparseSeries may lose dtype (GH13999)
  • Bug in SparseSeries comparison with dense returns normal Series rather than SparseSeries (GH13999)

Indexer dtype changes

Note

This change only affects 64 bit python running on Windows, and only affects relatively advanced indexing operations

Methods such as Index.get_indexer that return an indexer array, coerce that array to a “platform int”, so that it can be directly used in 3rd party library operations like numpy.take. Previously, a platform int was defined as np.int_ which corresponds to a C integer, but the correct type, and what is being used now, is np.intp, which corresponds to the C integer size that can hold a pointer (GH3033, GH13972).

These types are the same on many platform, but for 64 bit python on Windows, np.int_ is 32 bits, and np.intp is 64 bits. Changing this behavior improves performance for many operations on that platform.

Previous behavior:

In [1]: i = pd.Index(['a', 'b', 'c'])

In [2]: i.get_indexer(['b', 'b', 'c']).dtype
Out[2]: dtype('int32')

New behavior:

In [1]: i = pd.Index(['a', 'b', 'c'])

In [2]: i.get_indexer(['b', 'b', 'c']).dtype
Out[2]: dtype('int64')

Other API Changes

  • Timestamp.to_pydatetime will issue a UserWarning when warn=True, and the instance has a non-zero number of nanoseconds, previously this would print a message to stdout (GH14101).
  • Series.unique() with datetime and timezone now returns return array of Timestamp with timezone (GH13565).
  • Panel.to_sparse() will raise a NotImplementedError exception when called (GH13778).
  • Index.reshape() will raise a NotImplementedError exception when called (GH12882).
  • .filter() enforces mutual exclusion of the keyword arguments (GH12399).
  • eval‘s upcasting rules for float32 types have been updated to be more consistent with NumPy’s rules. New behavior will not upcast to float64 if you multiply a pandas float32 object by a scalar float64 (GH12388).
  • An UnsupportedFunctionCall error is now raised if NumPy ufuncs like np.mean are called on groupby or resample objects (GH12811).
  • __setitem__ will no longer apply a callable rhs as a function instead of storing it. Call where directly to get the previous behavior (GH13299).
  • Calls to .sample() will respect the random seed set via numpy.random.seed(n) (GH13161)
  • Styler.apply is now more strict about the outputs your function must return. For axis=0 or axis=1, the output shape must be identical. For axis=None, the output must be a DataFrame with identical columns and index labels (GH13222).
  • Float64Index.astype(int) will now raise ValueError if Float64Index contains NaN values (GH13149)
  • TimedeltaIndex.astype(int) and DatetimeIndex.astype(int) will now return Int64Index instead of np.array (GH13209)
  • Passing Period with multiple frequencies to normal Index now returns Index with object dtype (GH13664)
  • PeriodIndex.fillna with Period has different freq now coerces to object dtype (GH13664)
  • Faceted boxplots from DataFrame.boxplot(by=col) now return a Series when return_type is not None. Previously these returned an OrderedDict. Note that when return_type=None, the default, these still return a 2-D NumPy array (GH12216, GH7096).
  • pd.read_hdf will now raise a ValueError instead of KeyError, if a mode other than r, r+ and a is supplied. (GH13623)
  • pd.read_csv(), pd.read_table(), and pd.read_hdf() raise the builtin FileNotFoundError exception for Python 3.x when called on a nonexistent file; this is back-ported as IOError in Python 2.x (GH14086)
  • More informative exceptions are passed through the csv parser. The exception type would now be the original exception type instead of CParserError (GH13652).
  • pd.read_csv() in the C engine will now issue a ParserWarning or raise a ValueError when sep encoded is more than one character long (GH14065)
  • DataFrame.values will now return float64 with a DataFrame of mixed int64 and uint64 dtypes, conforming to np.find_common_type (GH10364, GH13917)
  • .groupby.groups will now return a dictionary of Index objects, rather than a dictionary of np.ndarray or lists (GH14293)

Deprecations

  • Series.reshape and Categorical.reshape have been deprecated and will be removed in a subsequent release (GH12882, GH12882)
  • PeriodIndex.to_datetime has been deprecated in favor of PeriodIndex.to_timestamp (GH8254)
  • Timestamp.to_datetime has been deprecated in favor of Timestamp.to_pydatetime (GH8254)
  • Index.to_datetime and DatetimeIndex.to_datetime have been deprecated in favor of pd.to_datetime (GH8254)
  • pandas.core.datetools module has been deprecated and will be removed in a subsequent release (GH14094)
  • SparseList has been deprecated and will be removed in a future version (GH13784)
  • DataFrame.to_html() and DataFrame.to_latex() have dropped the colSpace parameter in favor of col_space (GH13857)
  • DataFrame.to_sql() has deprecated the flavor parameter, as it is superfluous when SQLAlchemy is not installed (GH13611)
  • Deprecated read_csv keywords:
    • compact_ints and use_unsigned have been deprecated and will be removed in a future version (GH13320)
    • buffer_lines has been deprecated and will be removed in a future version (GH13360)
    • as_recarray has been deprecated and will be removed in a future version (GH13373)
    • skip_footer has been deprecated in favor of skipfooter and will be removed in a future version (GH13349)
  • top-level pd.ordered_merge() has been renamed to pd.merge_ordered() and the original name will be removed in a future version (GH13358)
  • Timestamp.offset property (and named arg in the constructor), has been deprecated in favor of freq (GH12160)
  • pd.tseries.util.pivot_annual is deprecated. Use pivot_table as alternative, an example is here (GH736)
  • pd.tseries.util.isleapyear has been deprecated and will be removed in a subsequent release. Datetime-likes now have a .is_leap_year property (GH13727)
  • Panel4D and PanelND constructors are deprecated and will be removed in a future version. The recommended way to represent these types of n-dimensional data are with the xarray package. Pandas provides a to_xarray() method to automate this conversion (GH13564).
  • pandas.tseries.frequencies.get_standard_freq is deprecated. Use pandas.tseries.frequencies.to_offset(freq).rule_code instead (GH13874)
  • pandas.tseries.frequencies.to_offset‘s freqstr keyword is deprecated in favor of freq (GH13874)
  • Categorical.from_array has been deprecated and will be removed in a future version (GH13854)

Removal of prior version deprecations/changes

  • The SparsePanel class has been removed (GH13778)
  • The pd.sandbox module has been removed in favor of the external library pandas-qt (GH13670)
  • The pandas.io.data and pandas.io.wb modules are removed in favor of the pandas-datareader package (GH13724).
  • The pandas.tools.rplot module has been removed in favor of the seaborn package (GH13855)
  • DataFrame.to_csv() has dropped the engine parameter, as was deprecated in 0.17.1 (GH11274, GH13419)
  • DataFrame.to_dict() has dropped the outtype parameter in favor of orient (GH13627, GH8486)
  • pd.Categorical has dropped setting of the ordered attribute directly in favor of the set_ordered method (GH13671)
  • pd.Categorical has dropped the levels attribute in favor of categories (GH8376)
  • DataFrame.to_sql() has dropped the mysql option for the flavor parameter (GH13611)
  • Panel.shift() has dropped the lags parameter in favor of periods (GH14041)
  • pd.Index has dropped the diff method in favor of difference (GH13669)
  • pd.DataFrame has dropped the to_wide method in favor of to_panel (GH14039)
  • Series.to_csv has dropped the nanRep parameter in favor of na_rep (GH13804)
  • Series.xs, DataFrame.xs, Panel.xs, Panel.major_xs, and Panel.minor_xs have dropped the copy parameter (GH13781)
  • str.split has dropped the return_type parameter in favor of expand (GH13701)
  • Removal of the legacy time rules (offset aliases), deprecated since 0.17.0 (this has been alias since 0.8.0) (GH13590, GH13868). Now legacy time rules raises ValueError. For the list of currently supported offsets, see here.
  • The default value for the return_type parameter for DataFrame.plot.box and DataFrame.boxplot changed from None to "axes". These methods will now return a matplotlib axes by default instead of a dictionary of artists. See here (GH6581).
  • The tquery and uquery functions in the pandas.io.sql module are removed (GH5950).

Performance Improvements

  • Improved performance of sparse IntIndex.intersect (GH13082)
  • Improved performance of sparse arithmetic with BlockIndex when the number of blocks are large, though recommended to use IntIndex in such cases (GH13082)
  • Improved performance of DataFrame.quantile() as it now operates per-block (GH11623)
  • Improved performance of float64 hash table operations, fixing some very slow indexing and groupby operations in python 3 (GH13166, GH13334)
  • Improved performance of DataFrameGroupBy.transform (GH12737)
  • Improved performance of Index and Series .duplicated (GH10235)
  • Improved performance of Index.difference (GH12044)
  • Improved performance of RangeIndex.is_monotonic_increasing and is_monotonic_decreasing (GH13749)
  • Improved performance of datetime string parsing in DatetimeIndex (GH13692)
  • Improved performance of hashing Period (GH12817)
  • Improved performance of factorize of datetime with timezone (GH13750)
  • Improved performance of by lazily creating indexing hashtables on larger Indexes (GH14266)
  • Improved performance of groupby.groups (GH14293)
  • Unecessary materializing of a MultiIndex when introspecting for memory usage (GH14308)

Bug Fixes

  • Bug in groupby().shift(), which could cause a segfault or corruption in rare circumstances when grouping by columns with missing values (GH13813)
  • Bug in groupby().cumsum() calculating cumprod when axis=1. (GH13994)
  • Bug in pd.to_timedelta() in which the errors parameter was not being respected (GH13613)
  • Bug in io.json.json_normalize(), where non-ascii keys raised an exception (GH13213)
  • Bug when passing a not-default-indexed Series as xerr or yerr in .plot() (GH11858)
  • Bug in area plot draws legend incorrectly if subplot is enabled or legend is moved after plot (matplotlib 1.5.0 is required to draw area plot legend properly) (GH9161, GH13544)
  • Bug in DataFrame assignment with an object-dtyped Index where the resultant column is mutable to the original object. (GH13522)
  • Bug in matplotlib AutoDataFormatter; this restores the second scaled formatting and re-adds micro-second scaled formatting (GH13131)
  • Bug in selection from a HDFStore with a fixed format and start and/or stop specified will now return the selected range (GH8287)
  • Bug in Categorical.from_codes() where an unhelpful error was raised when an invalid ordered parameter was passed in (GH14058)
  • Bug in Series construction from a tuple of integers on windows not returning default dtype (int64) (GH13646)
  • Bug in TimedeltaIndex addition with a Datetime-like object where addition overflow was not being caught (GH14068)
  • Bug in .groupby(..).resample(..) when the same object is called multiple times (GH13174)
  • Bug in .to_records() when index name is a unicode string (GH13172)
  • Bug in calling .memory_usage() on object which doesn’t implement (GH12924)
  • Regression in Series.quantile with nans (also shows up in .median() and .describe() ); furthermore now names the Series with the quantile (GH13098, GH13146)
  • Bug in SeriesGroupBy.transform with datetime values and missing groups (GH13191)
  • Bug where empty Series were incorrectly coerced in datetime-like numeric operations (GH13844)
  • Bug in Categorical constructor when passed a Categorical containing datetimes with timezones (GH14190)
  • Bug in Series.str.extractall() with str index raises ValueError (GH13156)
  • Bug in Series.str.extractall() with single group and quantifier (GH13382)
  • Bug in DatetimeIndex and Period subtraction raises ValueError or AttributeError rather than TypeError (GH13078)
  • Bug in Index and Series created with NaN and NaT mixed data may not have datetime64 dtype (GH13324)
  • Bug in Index and Series may ignore np.datetime64('nat') and np.timdelta64('nat') to infer dtype (GH13324)
  • Bug in PeriodIndex and Period subtraction raises AttributeError (GH13071)
  • Bug in PeriodIndex construction returning a float64 index in some circumstances (GH13067)
  • Bug in .resample(..) with a PeriodIndex not changing its freq appropriately when empty (GH13067)
  • Bug in .resample(..) with a PeriodIndex not retaining its type or name with an empty DataFrame appropriately when empty (GH13212)
  • Bug in groupby(..).apply(..) when the passed function returns scalar values per group (GH13468).
  • Bug in groupby(..).resample(..) where passing some keywords would raise an exception (GH13235)
  • Bug in .tz_convert on a tz-aware DateTimeIndex that relied on index being sorted for correct results (GH13306)
  • Bug in .tz_localize with dateutil.tz.tzlocal may return incorrect result (GH13583)
  • Bug in DatetimeTZDtype dtype with dateutil.tz.tzlocal cannot be regarded as valid dtype (GH13583)
  • Bug in pd.read_hdf() where attempting to load an HDF file with a single dataset, that had one or more categorical columns, failed unless the key argument was set to the name of the dataset. (GH13231)
  • Bug in .rolling() that allowed a negative integer window in contruction of the Rolling() object, but would later fail on aggregation (GH13383)
  • Bug in Series indexing with tuple-valued data and a numeric index (GH13509)
  • Bug in printing pd.DataFrame where unusual elements with the object dtype were causing segfaults (GH13717)
  • Bug in ranking Series which could result in segfaults (GH13445)
  • Bug in various index types, which did not propagate the name of passed index (GH12309)
  • Bug in DatetimeIndex, which did not honour the copy=True (GH13205)
  • Bug in DatetimeIndex.is_normalized returns incorrectly for normalized date_range in case of local timezones (GH13459)
  • Bug in pd.concat and .append may coerces datetime64 and timedelta to object dtype containing python built-in datetime or timedelta rather than Timestamp or Timedelta (GH13626)
  • Bug in PeriodIndex.append may raises AttributeError when the result is object dtype (GH13221)
  • Bug in CategoricalIndex.append may accept normal list (GH13626)
  • Bug in pd.concat and .append with the same timezone get reset to UTC (GH7795)
  • Bug in Series and DataFrame .append raises AmbiguousTimeError if data contains datetime near DST boundary (GH13626)
  • Bug in DataFrame.to_csv() in which float values were being quoted even though quotations were specified for non-numeric values only (GH12922, GH13259)
  • Bug in DataFrame.describe() raising ValueError with only boolean columns (GH13898)
  • Bug in MultiIndex slicing where extra elements were returned when level is non-unique (GH12896)
  • Bug in .str.replace does not raise TypeError for invalid replacement (GH13438)
  • Bug in MultiIndex.from_arrays which didn’t check for input array lengths matching (GH13599)
  • Bug in cartesian_product and MultiIndex.from_product which may raise with empty input arrays (GH12258)
  • Bug in pd.read_csv() which may cause a segfault or corruption when iterating in large chunks over a stream/file under rare circumstances (GH13703)
  • Bug in pd.read_csv() which caused errors to be raised when a dictionary containing scalars is passed in for na_values (GH12224)
  • Bug in pd.read_csv() which caused BOM files to be incorrectly parsed by not ignoring the BOM (GH4793)
  • Bug in pd.read_csv() with engine='python' which raised errors when a numpy array was passed in for usecols (GH12546)
  • Bug in pd.read_csv() where the index columns were being incorrectly parsed when parsed as dates with a thousands parameter (GH14066)
  • Bug in pd.read_csv() with engine='python' in which NaN values weren’t being detected after data was converted to numeric values (GH13314)
  • Bug in pd.read_csv() in which the nrows argument was not properly validated for both engines (GH10476)
  • Bug in pd.read_csv() with engine='python' in which infinities of mixed-case forms were not being interpreted properly (GH13274)
  • Bug in pd.read_csv() with engine='python' in which trailing NaN values were not being parsed (GH13320)
  • Bug in pd.read_csv() with engine='python' when reading from a tempfile.TemporaryFile on Windows with Python 3 (GH13398)
  • Bug in pd.read_csv() that prevents usecols kwarg from accepting single-byte unicode strings (GH13219)
  • Bug in pd.read_csv() that prevents usecols from being an empty set (GH13402)
  • Bug in pd.read_csv() in the C engine where the NULL character was not being parsed as NULL (GH14012)
  • Bug in pd.read_csv() with engine='c' in which NULL quotechar was not accepted even though quoting was specified as None (GH13411)
  • Bug in pd.read_csv() with engine='c' in which fields were not properly cast to float when quoting was specified as non-numeric (GH13411)
  • Bug in pd.read_csv() in Python 2.x with non-UTF8 encoded, multi-character separated data (GH3404)
  • Bug in pd.read_csv(), where aliases for utf-xx (e.g. UTF-xx, UTF_xx, utf_xx) raised UnicodeDecodeError (GH13549)
  • Bug in pd.read_csv, pd.read_table, pd.read_fwf, pd.read_stata and pd.read_sas where files were opened by parsers but not closed if both chunksize and iterator were None. (GH13940)
  • Bug in StataReader, StataWriter, XportReader and SAS7BDATReader where a file was not properly closed when an error was raised. (GH13940)
  • Bug in pd.pivot_table() where margins_name is ignored when aggfunc is a list (GH13354)
  • Bug in pd.Series.str.zfill, center, ljust, rjust, and pad when passing non-integers, did not raise TypeError (GH13598)
  • Bug in checking for any null objects in a TimedeltaIndex, which always returned True (GH13603)
  • Bug in Series arithmetic raises TypeError if it contains datetime-like as object dtype (GH13043)
  • Bug Series.isnull() and Series.notnull() ignore Period('NaT') (GH13737)
  • Bug Series.fillna() and Series.dropna() don’t affect to Period('NaT') (GH13737
  • Bug in .fillna(value=np.nan) incorrectly raises KeyError on a category dtyped Series (GH14021)
  • Bug in extension dtype creation where the created types were not is/identical (GH13285)
  • Bug in .resample(..) where incorrect warnings were triggered by IPython introspection (GH13618)
  • Bug in NaT - Period raises AttributeError (GH13071)
  • Bug in Series comparison may output incorrect result if rhs contains NaT (GH9005)
  • Bug in Series and Index comparison may output incorrect result if it contains NaT with object dtype (GH13592)
  • Bug in Period addition raises TypeError if Period is on right hand side (GH13069)
  • Bug in Peirod and Series or Index comparison raises TypeError (GH13200)
  • Bug in pd.set_eng_float_format() that would prevent NaN and Inf from formatting (GH11981)
  • Bug in .unstack with Categorical dtype resets .ordered to True (GH13249)
  • Clean some compile time warnings in datetime parsing (GH13607)
  • Bug in factorize raises AmbiguousTimeError if data contains datetime near DST boundary (GH13750)
  • Bug in .set_index raises AmbiguousTimeError if new index contains DST boundary and multi levels (GH12920)
  • Bug in .shift raises AmbiguousTimeError if data contains datetime near DST boundary (GH13926)
  • Bug in pd.read_hdf() returns incorrect result when a DataFrame with a categorical column and a query which doesn’t match any values (GH13792)
  • Bug in .iloc when indexing with a non lex-sorted MultiIndex (GH13797)
  • Bug in .loc when indexing with date strings in a reverse sorted DatetimeIndex (GH14316)
  • Bug in Series comparison operators when dealing with zero dim NumPy arrays (GH13006)
  • Bug in .combine_first may return incorrect dtype (GH7630, GH10567)
  • Bug in groupby where apply returns different result depending on whether first result is None or not (GH12824)
  • Bug in groupby(..).nth() where the group key is included inconsistently if called after .head()/.tail() (GH12839)
  • Bug in .to_html, .to_latex and .to_string silently ignore custom datetime formatter passed through the formatters key word (GH10690)
  • Bug in DataFrame.iterrows(), not yielding a Series subclasse if defined (GH13977)
  • Bug in pd.to_numeric when errors='coerce' and input contains non-hashable objects (GH13324)
  • Bug in invalid Timedelta arithmetic and comparison may raise ValueError rather than TypeError (GH13624)
  • Bug in invalid datetime parsing in to_datetime and DatetimeIndex may raise TypeError rather than ValueError (GH11169, GH11287)
  • Bug in Index created with tz-aware Timestamp and mismatched tz option incorrectly coerces timezone (GH13692)
  • Bug in DatetimeIndex with nanosecond frequency does not include timestamp specified with end (GH13672)
  • Bug in `Series` when setting a slice with a `np.timedelta64` (GH14155)
  • Bug in Index raises OutOfBoundsDatetime if datetime exceeds datetime64[ns] bounds, rather than coercing to object dtype (GH13663)
  • Bug in Index may ignore specified datetime64 or timedelta64 passed as dtype (GH13981)
  • Bug in RangeIndex can be created without no arguments rather than raises TypeError (GH13793)
  • Bug in .value_counts() raises OutOfBoundsDatetime if data exceeds datetime64[ns] bounds (GH13663)
  • Bug in DatetimeIndex may raise OutOfBoundsDatetime if input np.datetime64 has other unit than ns (GH9114)
  • Bug in Series creation with np.datetime64 which has other unit than ns as object dtype results in incorrect values (GH13876)
  • Bug in resample with timedelta data where data was casted to float (GH13119).
  • Bug in pd.isnull() pd.notnull() raise TypeError if input datetime-like has other unit than ns (GH13389)
  • Bug in pd.merge() may raise TypeError if input datetime-like has other unit than ns (GH13389)
  • Bug in HDFStore/read_hdf() discarded DatetimeIndex.name if tz was set (GH13884)
  • Bug in Categorical.remove_unused_categories() changes .codes dtype to platform int (GH13261)
  • Bug in groupby with as_index=False returns all NaN’s when grouping on multiple columns including a categorical one (GH13204)
  • Bug in df.groupby(...)[...] where getitem with Int64Index raised an error (GH13731)
  • Bug in the CSS classes assigned to DataFrame.style for index names. Previously they were assigned "col_heading level<n> col<c>" where n was the number of levels + 1. Now they are assigned "index_name level<n>", where n is the correct level for that MultiIndex.
  • Bug where pd.read_gbq() could throw ImportError: No module named discovery as a result of a naming conflict with another python package called apiclient (GH13454)
  • Bug in Index.union returns an incorrect result with a named empty index (GH13432)
  • Bugs in Index.difference and DataFrame.join raise in Python3 when using mixed-integer indexes (GH13432, GH12814)
  • Bug in subtract tz-aware datetime.datetime from tz-aware datetime64 series (GH14088)
  • Bug in .to_excel() when DataFrame contains a MultiIndex which contains a label with a NaN value (GH13511)
  • Bug in invalid frequency offset string like “D1”, “-2-3H” may not raise ValueError (GH13930)
  • Bug in concat and groupby for hierarchical frames with RangeIndex levels (GH13542).
  • Bug in Series.str.contains() for Series containing only NaN values of object dtype (GH14171)
  • Bug in agg() function on groupby dataframe changes dtype of datetime64[ns] column to float64 (GH12821)
  • Bug in using NumPy ufunc with PeriodIndex to add or subtract integer raise IncompatibleFrequency. Note that using standard operator like + or - is recommended, because standard operators use more efficient path (GH13980)
  • Bug in operations on NaT returning float instead of datetime64[ns] (GH12941)
  • Bug in Series flexible arithmetic methods (like .add()) raises ValueError when axis=None (GH13894)
  • Bug in DataFrame.to_csv() with MultiIndex columns in which a stray empty line was added (GH6618)
  • Bug in DatetimeIndex, TimedeltaIndex and PeriodIndex.equals() may return True when input isn’t Index but contains the same values (GH13107)
  • Bug in assignment against datetime with timezone may not work if it contains datetime near DST boundary (GH14146)
  • Bug in pd.eval() and HDFStore query truncating long float literals with python 2 (GH14241)
  • Bug in Index raises KeyError displaying incorrect column when column is not in the df and columns contains duplicate values (GH13822)
  • Bug in Period and PeriodIndex creating wrong dates when frequency has combined offset aliases (GH13874)
  • Bug in .to_string() when called with an integer line_width and index=False raises an UnboundLocalError exception because idx referenced before assignment.
  • Bug in eval() where the resolvers argument would not accept a list (GH14095)
  • Bugs in stack, get_dummies, make_axis_dummies which don’t preserve categorical dtypes in (multi)indexes (GH13854)
  • PeriodIndex can now accept list and array which contains pd.NaT (GH13430)
  • Bug in df.groupby where .median() returns arbitrary values if grouped dataframe contains empty bins (GH13629)
  • Bug in Index.copy() where name parameter was ignored (GH14302)

v0.18.1 (May 3, 2016)

This is a minor bug-fix release from 0.18.0 and includes a large number of bug fixes along with several new features, enhancements, and performance improvements. We recommend that all users upgrade to this version.

Highlights include:

  • .groupby(...) has been enhanced to provide convenient syntax when working with .rolling(..), .expanding(..) and .resample(..) per group, see here
  • pd.to_datetime() has gained the ability to assemble dates from a DataFrame, see here
  • Method chaining improvements, see here.
  • Custom business hour offset, see here.
  • Many bug fixes in the handling of sparse, see here
  • Expanded the Tutorials section with a feature on modern pandas, courtesy of @TomAugsburger. (GH13045).

New features

Custom Business Hour

The CustomBusinessHour is a mixture of BusinessHour and CustomBusinessDay which allows you to specify arbitrary holidays. For details, see Custom Business Hour (GH11514)

In [1]: from pandas.tseries.offsets import CustomBusinessHour

In [2]: from pandas.tseries.holiday import USFederalHolidayCalendar

In [3]: bhour_us = CustomBusinessHour(calendar=USFederalHolidayCalendar())

Friday before MLK Day

In [4]: dt = datetime(2014, 1, 17, 15)

In [5]: dt + bhour_us
Out[5]: Timestamp('2014-01-17 16:00:00')

Tuesday after MLK Day (Monday is skipped because it’s a holiday)

In [6]: dt + bhour_us * 2
Out[6]: Timestamp('2014-01-20 09:00:00')

.groupby(..) syntax with window and resample operations

.groupby(...) has been enhanced to provide convenient syntax when working with .rolling(..), .expanding(..) and .resample(..) per group, see (GH12486, GH12738).

You can now use .rolling(..) and .expanding(..) as methods on groupbys. These return another deferred object (similar to what .rolling() and .expanding() do on ungrouped pandas objects). You can then operate on these RollingGroupby objects in a similar manner.

Previously you would have to do this to get a rolling window mean per-group:

In [7]: df = pd.DataFrame({'A': [1] * 20 + [2] * 12 + [3] * 8,
   ...:                    'B': np.arange(40)})
   ...: 

In [8]: df
Out[8]: 
    A   B
0   1   0
1   1   1
2   1   2
3   1   3
4   1   4
5   1   5
6   1   6
.. ..  ..
33  3  33
34  3  34
35  3  35
36  3  36
37  3  37
38  3  38
39  3  39

[40 rows x 2 columns]
In [9]: df.groupby('A').apply(lambda x: x.rolling(4).B.mean())
Out[9]: 
A    
1  0      NaN
   1      NaN
   2      NaN
   3      1.5
   4      2.5
   5      3.5
   6      4.5
         ... 
3  33     NaN
   34     NaN
   35    33.5
   36    34.5
   37    35.5
   38    36.5
   39    37.5
Name: B, dtype: float64

Now you can do:

In [10]: df.groupby('A').rolling(4).B.mean()
Out[10]: 
A    
1  0      NaN
   1      NaN
   2      NaN
   3      1.5
   4      2.5
   5      3.5
   6      4.5
         ... 
3  33     NaN
   34     NaN
   35    33.5
   36    34.5
   37    35.5
   38    36.5
   39    37.5
Name: B, dtype: float64

For .resample(..) type of operations, previously you would have to:

In [11]: df = pd.DataFrame({'date': pd.date_range(start='2016-01-01',
   ....:                                          periods=4,
   ....:                                          freq='W'),
   ....:                    'group': [1, 1, 2, 2],
   ....:                    'val': [5, 6, 7, 8]}).set_index('date')
   ....: 

In [12]: df
Out[12]: 
            group  val
date                  
2016-01-03      1    5
2016-01-10      1    6
2016-01-17      2    7
2016-01-24      2    8
In [13]: df.groupby('group').apply(lambda x: x.resample('1D').ffill())
Out[13]: 
                  group  val
group date                  
1     2016-01-03      1    5
      2016-01-04      1    5
      2016-01-05      1    5
      2016-01-06      1    5
      2016-01-07      1    5
      2016-01-08      1    5
      2016-01-09      1    5
...                 ...  ...
2     2016-01-18      2    7
      2016-01-19      2    7
      2016-01-20      2    7
      2016-01-21      2    7
      2016-01-22      2    7
      2016-01-23      2    7
      2016-01-24      2    8

[16 rows x 2 columns]

Now you can do:

In [14]: df.groupby('group').resample('1D').ffill()
Out[14]: 
                  group  val
group date                  
1     2016-01-03      1    5
      2016-01-04      1    5
      2016-01-05      1    5
      2016-01-06      1    5
      2016-01-07      1    5
      2016-01-08      1    5
      2016-01-09      1    5
...                 ...  ...
2     2016-01-18      2    7
      2016-01-19      2    7
      2016-01-20      2    7
      2016-01-21      2    7
      2016-01-22      2    7
      2016-01-23      2    7
      2016-01-24      2    8

[16 rows x 2 columns]

Method chaininng improvements

The following methods / indexers now accept a callable. It is intended to make these more useful in method chains, see the documentation. (GH11485, GH12533)

  • .where() and .mask()
  • .loc[], iloc[] and .ix[]
  • [] indexing
.where() and .mask()

These can accept a callable for the condition and other arguments.

In [15]: df = pd.DataFrame({'A': [1, 2, 3],
   ....:                    'B': [4, 5, 6],
   ....:                    'C': [7, 8, 9]})
   ....: 

In [16]: df.where(lambda x: x > 4, lambda x: x + 10)
Out[16]: 
    A   B  C
0  11  14  7
1  12   5  8
2  13   6  9
.loc[], .iloc[], .ix[]

These can accept a callable, and a tuple of callable as a slicer. The callable can return a valid boolean indexer or anything which is valid for these indexer’s input.

# callable returns bool indexer
In [17]: df.loc[lambda x: x.A >= 2, lambda x: x.sum() > 10]
Out[17]: 
   B  C
1  5  8
2  6  9

# callable returns list of labels
In [18]: df.loc[lambda x: [1, 2], lambda x: ['A', 'B']]
Out[18]: 
   A  B
1  2  5
2  3  6
[] indexing

Finally, you can use a callable in [] indexing of Series, DataFrame and Panel. The callable must return a valid input for [] indexing depending on its class and index type.

In [19]: df[lambda x: 'A']
Out[19]: 
0    1
1    2
2    3
Name: A, dtype: int64

Using these methods / indexers, you can chain data selection operations without using temporary variable.

In [20]: bb = pd.read_csv('data/baseball.csv', index_col='id')

In [21]: (bb.groupby(['year', 'team'])
   ....:    .sum()
   ....:    .loc[lambda df: df.r > 100]
   ....: )
   ....: 
Out[21]: 
           stint    g    ab    r    h  X2b  X3b  hr    rbi    sb   cs   bb  \
year team                                                                    
2007 CIN       6  379   745  101  203   35    2  36  125.0  10.0  1.0  105   
     DET       5  301  1062  162  283   54    4  37  144.0  24.0  7.0   97   
     HOU       4  311   926  109  218   47    6  14   77.0  10.0  4.0   60   
     LAN      11  413  1021  153  293   61    3  36  154.0   7.0  5.0  114   
     NYN      13  622  1854  240  509  101    3  61  243.0  22.0  4.0  174   
     SFN       5  482  1305  198  337   67    6  40  171.0  26.0  7.0  235   
     TEX       2  198   729  115  200   40    4  28  115.0  21.0  4.0   73   
     TOR       4  459  1408  187  378   96    2  58  223.0   4.0  2.0  190   

              so   ibb   hbp    sh    sf  gidp  
year team                                       
2007 CIN   127.0  14.0   1.0   1.0  15.0  18.0  
     DET   176.0   3.0  10.0   4.0   8.0  28.0  
     HOU   212.0   3.0   9.0  16.0   6.0  17.0  
     LAN   141.0   8.0   9.0   3.0   8.0  29.0  
     NYN   310.0  24.0  23.0  18.0  15.0  48.0  
     SFN   188.0  51.0   8.0  16.0   6.0  41.0  
     TEX   140.0   4.0   5.0   2.0   8.0  16.0  
     TOR   265.0  16.0  12.0   4.0  16.0  38.0  

Partial string indexing on DateTimeIndex when part of a MultiIndex

Partial string indexing now matches on DateTimeIndex when part of a MultiIndex (GH10331)

In [22]: dft2 = pd.DataFrame(np.random.randn(20, 1),
   ....:                     columns=['A'],
   ....:                     index=pd.MultiIndex.from_product([pd.date_range('20130101',
   ....:                                                                     periods=10,
   ....:                                                                     freq='12H'),
   ....:                                                      ['a', 'b']]))
   ....: 

In [23]: dft2
Out[23]: 
                              A
2013-01-01 00:00:00 a  1.474071
                    b -0.064034
2013-01-01 12:00:00 a -1.282782
                    b  0.781836
2013-01-02 00:00:00 a -1.071357
                    b  0.441153
2013-01-02 12:00:00 a  2.353925
...                         ...
2013-01-04 00:00:00 b -0.845696
2013-01-04 12:00:00 a -1.340896
                    b  1.846883
2013-01-05 00:00:00 a -1.328865
                    b  1.682706
2013-01-05 12:00:00 a -1.717693
                    b  0.888782

[20 rows x 1 columns]

In [24]: dft2.loc['2013-01-05']
Out[24]: 
                              A
2013-01-05 00:00:00 a -1.328865
                    b  1.682706
2013-01-05 12:00:00 a -1.717693
                    b  0.888782

On other levels

In [25]: idx = pd.IndexSlice

In [26]: dft2 = dft2.swaplevel(0, 1).sort_index()

In [27]: dft2
Out[27]: 
                              A
a 2013-01-01 00:00:00  1.474071
  2013-01-01 12:00:00 -1.282782
  2013-01-02 00:00:00 -1.071357
  2013-01-02 12:00:00  2.353925
  2013-01-03 00:00:00  0.221471
  2013-01-03 12:00:00  0.758527
  2013-01-04 00:00:00 -0.964980
...                         ...
b 2013-01-02 12:00:00  0.583787
  2013-01-03 00:00:00 -0.744471
  2013-01-03 12:00:00  1.729689
  2013-01-04 00:00:00 -0.845696
  2013-01-04 12:00:00  1.846883
  2013-01-05 00:00:00  1.682706
  2013-01-05 12:00:00  0.888782

[20 rows x 1 columns]

In [28]: dft2.loc[idx[:, '2013-01-05'], :]
Out[28]: 
                              A
a 2013-01-05 00:00:00 -1.328865
  2013-01-05 12:00:00 -1.717693
b 2013-01-05 00:00:00  1.682706
  2013-01-05 12:00:00  0.888782

Assembling Datetimes

pd.to_datetime() has gained the ability to assemble datetimes from a passed in DataFrame or a dict. (GH8158).

In [29]: df = pd.DataFrame({'year': [2015, 2016],
   ....:                    'month': [2, 3],
   ....:                    'day': [4, 5],
   ....:                    'hour': [2, 3]})
   ....: 

In [30]: df
Out[30]: 
   day  hour  month  year
0    4     2      2  2015
1    5     3      3  2016

Assembling using the passed frame.

In [31]: pd.to_datetime(df)
Out[31]: 
0   2015-02-04 02:00:00
1   2016-03-05 03:00:00
dtype: datetime64[ns]

You can pass only the columns that you need to assemble.

In [32]: pd.to_datetime(df[['year', 'month', 'day']])
Out[32]: 
0   2015-02-04
1   2016-03-05
dtype: datetime64[ns]

Other Enhancements

  • pd.read_csv() now supports delim_whitespace=True for the Python engine (GH12958)

  • pd.read_csv() now supports opening ZIP files that contains a single CSV, via extension inference or explict compression='zip' (GH12175)

  • pd.read_csv() now supports opening files using xz compression, via extension inference or explicit compression='xz' is specified; xz compressions is also supported by DataFrame.to_csv in the same way (GH11852)

  • pd.read_msgpack() now always gives writeable ndarrays even when compression is used (GH12359).

  • pd.read_msgpack() now supports serializing and de-serializing categoricals with msgpack (GH12573)

  • .to_json() now supports NDFrames that contain categorical and sparse data (GH10778)

  • interpolate() now supports method='akima' (GH7588).

  • pd.read_excel() now accepts path objects (e.g. pathlib.Path, py.path.local) for the file path, in line with other read_* functions (GH12655)

  • Added .weekday_name property as a component to DatetimeIndex and the .dt accessor. (GH11128)

  • Index.take now handles allow_fill and fill_value consistently (GH12631)

    In [33]: idx = pd.Index([1., 2., 3., 4.], dtype='float')
    
    # default, allow_fill=True, fill_value=None
    In [34]: idx.take([2, -1])
    Out[34]: Float64Index([3.0, 4.0], dtype='float64')
    
    In [35]: idx.take([2, -1], fill_value=True)
    Out[35]: Float64Index([3.0, nan], dtype='float64')
    
  • Index now supports .str.get_dummies() which returns MultiIndex, see Creating Indicator Variables (GH10008, GH10103)

    In [36]: idx = pd.Index(['a|b', 'a|c', 'b|c'])
    
    In [37]: idx.str.get_dummies('|')
    Out[37]: 
    MultiIndex(levels=[[0, 1], [0, 1], [0, 1]],
               labels=[[1, 1, 0], [1, 0, 1], [0, 1, 1]],
               names=[u'a', u'b', u'c'])
    
  • pd.crosstab() has gained a normalize argument for normalizing frequency tables (GH12569). Examples in the updated docs here.

  • .resample(..).interpolate() is now supported (GH12925)

  • .isin() now accepts passed sets (GH12988)

Sparse changes

These changes conform sparse handling to return the correct types and work to make a smoother experience with indexing.

SparseArray.take now returns a scalar for scalar input, SparseArray for others. Furthermore, it handles a negative indexer with the same rule as Index (GH10560, GH12796)

In [38]: s = pd.SparseArray([np.nan, np.nan, 1, 2, 3, np.nan, 4, 5, np.nan, 6])

In [39]: s.take(0)
Out[39]: nan

In [40]: s.take([1, 2, 3])
Out[40]: 
[nan, 1.0, 2.0]
Fill: nan
IntIndex
Indices: array([1, 2], dtype=int32)
  • Bug in SparseSeries[] indexing with Ellipsis raises KeyError (GH9467)
  • Bug in SparseArray[] indexing with tuples are not handled properly (GH12966)
  • Bug in SparseSeries.loc[] with list-like input raises TypeError (GH10560)
  • Bug in SparseSeries.iloc[] with scalar input may raise IndexError (GH10560)
  • Bug in SparseSeries.loc[], .iloc[] with slice returns SparseArray, rather than SparseSeries (GH10560)
  • Bug in SparseDataFrame.loc[], .iloc[] may results in dense Series, rather than SparseSeries (GH12787)
  • Bug in SparseArray addition ignores fill_value of right hand side (GH12910)
  • Bug in SparseArray mod raises AttributeError (GH12910)
  • Bug in SparseArray pow calculates 1 ** np.nan as np.nan which must be 1 (GH12910)
  • Bug in SparseArray comparison output may incorrect result or raise ValueError (GH12971)
  • Bug in SparseSeries.__repr__ raises TypeError when it is longer than max_rows (GH10560)
  • Bug in SparseSeries.shape ignores fill_value (GH10452)
  • Bug in SparseSeries and SparseArray may have different dtype from its dense values (GH12908)
  • Bug in SparseSeries.reindex incorrectly handle fill_value (GH12797)
  • Bug in SparseArray.to_frame() results in DataFrame, rather than SparseDataFrame (GH9850)
  • Bug in SparseSeries.value_counts() does not count fill_value (GH6749)
  • Bug in SparseArray.to_dense() does not preserve dtype (GH10648)
  • Bug in SparseArray.to_dense() incorrectly handle fill_value (GH12797)
  • Bug in pd.concat() of SparseSeries results in dense (GH10536)
  • Bug in pd.concat() of SparseDataFrame incorrectly handle fill_value (GH9765)
  • Bug in pd.concat() of SparseDataFrame may raise AttributeError (GH12174)
  • Bug in SparseArray.shift() may raise NameError or TypeError (GH12908)

API changes

.groupby(..).nth() changes

The index in .groupby(..).nth() output is now more consistent when the as_index argument is passed (GH11039):

In [41]: df = DataFrame({'A' : ['a', 'b', 'a'],
   ....:                 'B' : [1, 2, 3]})
   ....: 

In [42]: df
Out[42]: 
   A  B
0  a  1
1  b  2
2  a  3

Previous Behavior:

In [3]: df.groupby('A', as_index=True)['B'].nth(0)
Out[3]:
0    1
1    2
Name: B, dtype: int64

In [4]: df.groupby('A', as_index=False)['B'].nth(0)
Out[4]:
0    1
1    2
Name: B, dtype: int64

New Behavior:

In [43]: df.groupby('A', as_index=True)['B'].nth(0)
Out[43]: 
A
a    1
b    2
Name: B, dtype: int64

In [44]: df.groupby('A', as_index=False)['B'].nth(0)
Out[44]: 
0    1
1    2
Name: B, dtype: int64

Furthermore, previously, a .groupby would always sort, regardless if sort=False was passed with .nth().

In [45]: np.random.seed(1234)

In [46]: df = pd.DataFrame(np.random.randn(100, 2), columns=['a', 'b'])

In [47]: df['c'] = np.random.randint(0, 4, 100)

Previous Behavior:

In [4]: df.groupby('c', sort=True).nth(1)
Out[4]:
          a         b
c
0 -0.334077  0.002118
1  0.036142 -2.074978
2 -0.720589  0.887163
3  0.859588 -0.636524

In [5]: df.groupby('c', sort=False).nth(1)
Out[5]:
          a         b
c
0 -0.334077  0.002118
1  0.036142 -2.074978
2 -0.720589  0.887163
3  0.859588 -0.636524

New Behavior:

In [48]: df.groupby('c', sort=True).nth(1)
Out[48]: 
          a         b
c                    
0 -0.334077  0.002118
1  0.036142 -2.074978
2 -0.720589  0.887163
3  0.859588 -0.636524

In [49]: df.groupby('c', sort=False).nth(1)
Out[49]: 
          a         b
c                    
2 -0.720589  0.887163
3  0.859588 -0.636524
0 -0.334077  0.002118
1  0.036142 -2.074978

numpy function compatibility

Compatibility between pandas array-like methods (e.g. sum and take) and their numpy counterparts has been greatly increased by augmenting the signatures of the pandas methods so as to accept arguments that can be passed in from numpy, even if they are not necessarily used in the pandas implementation (GH12644, GH12638, GH12687)

  • .searchsorted() for Index and TimedeltaIndex now accept a sorter argument to maintain compatibility with numpy’s searchsorted function (GH12238)
  • Bug in numpy compatibility of np.round() on a Series (GH12600)

An example of this signature augmentation is illustrated below:

In [50]: sp = pd.SparseDataFrame([1, 2, 3])

In [51]: sp
Out[51]: 
   0
0  1
1  2
2  3

Previous behaviour:

In [2]: np.cumsum(sp, axis=0)
...
TypeError: cumsum() takes at most 2 arguments (4 given)

New behaviour:

In [52]: np.cumsum(sp, axis=0)
Out[52]: 
   0
0  1
1  3
2  6

Using .apply on groupby resampling

Using apply on resampling groupby operations (using a pd.TimeGrouper) now has the same output types as similar apply calls on other groupby operations. (GH11742).

In [53]: df = pd.DataFrame({'date': pd.to_datetime(['10/10/2000', '11/10/2000']),
   ....:                   'value': [10, 13]})
   ....: 

In [54]: df
Out[54]: 
        date  value
0 2000-10-10     10
1 2000-11-10     13

Previous behavior:

In [1]: df.groupby(pd.TimeGrouper(key='date', freq='M')).apply(lambda x: x.value.sum())
Out[1]:
...
TypeError: cannot concatenate a non-NDFrame object

# Output is a Series
In [2]: df.groupby(pd.TimeGrouper(key='date', freq='M')).apply(lambda x: x[['value']].sum())
Out[2]:
date
2000-10-31  value    10
2000-11-30  value    13
dtype: int64

New Behavior:

# Output is a Series
In [55]: df.groupby(pd.TimeGrouper(key='date', freq='M')).apply(lambda x: x.value.sum())
Out[55]: 
date
2000-10-31    10
2000-11-30    13
Freq: M, dtype: int64

# Output is a DataFrame
In [56]: df.groupby(pd.TimeGrouper(key='date', freq='M')).apply(lambda x: x[['value']].sum())
Out[56]: 
            value
date             
2000-10-31     10
2000-11-30     13

Changes in read_csv exceptions

In order to standardize the read_csv API for both the c and python engines, both will now raise an EmptyDataError, a subclass of ValueError, in response to empty columns or header (GH12493, GH12506)

Previous behaviour:

In [1]: df = pd.read_csv(StringIO(''), engine='c')
...
ValueError: No columns to parse from file

In [2]: df = pd.read_csv(StringIO(''), engine='python')
...
StopIteration

New behaviour:

In [1]: df = pd.read_csv(StringIO(''), engine='c')
...
pandas.io.common.EmptyDataError: No columns to parse from file

In [2]: df = pd.read_csv(StringIO(''), engine='python')
...
pandas.io.common.EmptyDataError: No columns to parse from file

In addition to this error change, several others have been made as well:

  • CParserError now sub-classes ValueError instead of just a Exception (GH12551)
  • A CParserError is now raised instead of a generic Exception in read_csv when the c engine cannot parse a column (GH12506)
  • A ValueError is now raised instead of a generic Exception in read_csv when the c engine encounters a NaN value in an integer column (GH12506)
  • A ValueError is now raised instead of a generic Exception in read_csv when true_values is specified, and the c engine encounters an element in a column containing unencodable bytes (GH12506)
  • pandas.parser.OverflowError exception has been removed and has been replaced with Python’s built-in OverflowError exception (GH12506)
  • pd.read_csv() no longer allows a combination of strings and integers for the usecols parameter (GH12678)

to_datetime error changes

Bugs in pd.to_datetime() when passing a unit with convertible entries and errors='coerce' or non-convertible with errors='ignore'. Furthermore, an OutOfBoundsDateime exception will be raised when an out-of-range value is encountered for that unit when errors='raise'. (GH11758, GH13052, GH13059)

Previous behaviour:

In [27]: pd.to_datetime(1420043460, unit='s', errors='coerce')
Out[27]: NaT

In [28]: pd.to_datetime(11111111, unit='D', errors='ignore')
OverflowError: Python int too large to convert to C long

In [29]: pd.to_datetime(11111111, unit='D', errors='raise')
OverflowError: Python int too large to convert to C long

New behaviour:

In [2]: pd.to_datetime(1420043460, unit='s', errors='coerce')
Out[2]: Timestamp('2014-12-31 16:31:00')

In [3]: pd.to_datetime(11111111, unit='D', errors='ignore')
Out[3]: 11111111

In [4]: pd.to_datetime(11111111, unit='D', errors='raise')
OutOfBoundsDatetime: cannot convert input with unit 'D'

Other API changes

  • .swaplevel() for Series, DataFrame, Panel, and MultiIndex now features defaults for its first two parameters i and j that swap the two innermost levels of the index. (GH12934)
  • .searchsorted() for Index and TimedeltaIndex now accept a sorter argument to maintain compatibility with numpy’s searchsorted function (GH12238)
  • Period and PeriodIndex now raises IncompatibleFrequency error which inherits ValueError rather than raw ValueError (GH12615)
  • Series.apply for category dtype now applies the passed function to each of the .categories (and not the .codes), and returns a category dtype if possible (GH12473)
  • read_csv will now raise a TypeError if parse_dates is neither a boolean, list, or dictionary (matches the doc-string) (GH5636)
  • The default for .query()/.eval() is now engine=None, which will use numexpr if it’s installed; otherwise it will fallback to the python engine. This mimics the pre-0.18.1 behavior if numexpr is installed (and which, previously, if numexpr was not installed, .query()/.eval() would raise). (GH12749)
  • pd.show_versions() now includes pandas_datareader version (GH12740)
  • Provide a proper __name__ and __qualname__ attributes for generic functions (GH12021)
  • pd.concat(ignore_index=True) now uses RangeIndex as default (GH12695)
  • pd.merge() and DataFrame.join() will show a UserWarning when merging/joining a single- with a multi-leveled dataframe (GH9455, GH12219)
  • Compat with scipy > 0.17 for deprecated piecewise_polynomial interpolation method; support for the replacement from_derivatives method (GH12887)

Deprecations

  • The method name Index.sym_diff() is deprecated and can be replaced by Index.symmetric_difference() (GH12591)
  • The method name Categorical.sort() is deprecated in favor of Categorical.sort_values() (GH12882)

Performance Improvements

  • Improved speed of SAS reader (GH12656, GH12961)
  • Performance improvements in .groupby(..).cumcount() (GH11039)
  • Improved memory usage in pd.read_csv() when using skiprows=an_integer (GH13005)
  • Improved performance of DataFrame.to_sql when checking case sensitivity for tables. Now only checks if table has been created correctly when table name is not lower case. (GH12876)
  • Improved performance of Period construction and time series plotting (GH12903, GH11831).
  • Improved performance of .str.encode() and .str.decode() methods (GH13008)
  • Improved performance of to_numeric if input is numeric dtype (GH12777)
  • Improved performance of sparse arithmetic with IntIndex (GH13036)

Bug Fixes

  • usecols parameter in pd.read_csv is now respected even when the lines of a CSV file are not even (GH12203)
  • Bug in groupby.transform(..) when axis=1 is specified with a non-monotonic ordered index (GH12713)
  • Bug in Period and PeriodIndex creation raises KeyError if freq="Minute" is specified. Note that “Minute” freq is deprecated in v0.17.0, and recommended to use freq="T" instead (GH11854)
  • Bug in .resample(...).count() with a PeriodIndex always raising a TypeError (GH12774)
  • Bug in .resample(...) with a PeriodIndex casting to a DatetimeIndex when empty (GH12868)
  • Bug in .resample(...) with a PeriodIndex when resampling to an existing frequency (GH12770)
  • Bug in printing data which contains Period with different freq raises ValueError (GH12615)
  • Bug in Series construction with Categorical and dtype='category' is specified (GH12574)
  • Bugs in concatenation with a coercable dtype was too aggressive, resulting in different dtypes in outputformatting when an object was longer than display.max_rows (GH12411, GH12045, GH11594, GH10571, GH12211)
  • Bug in float_format option with option not being validated as a callable. (GH12706)
  • Bug in GroupBy.filter when dropna=False and no groups fulfilled the criteria (GH12768)
  • Bug in __name__ of .cum* functions (GH12021)
  • Bug in .astype() of a Float64Inde/Int64Index to an Int64Index (GH12881)
  • Bug in roundtripping an integer based index in .to_json()/.read_json() when orient='index' (the default) (GH12866)
  • Bug in plotting Categorical dtypes cause error when attempting stacked bar plot (GH13019)
  • Compat with >= numpy 1.11 for NaT comparions (GH12969)
  • Bug in .drop() with a non-unique MultiIndex. (GH12701)
  • Bug in .concat of datetime tz-aware and naive DataFrames (GH12467)
  • Bug in correctly raising a ValueError in .resample(..).fillna(..) when passing a non-string (GH12952)
  • Bug fixes in various encoding and header processing issues in pd.read_sas() (GH12659, GH12654, GH12647, GH12809)
  • Bug in pd.crosstab() where would silently ignore aggfunc if values=None (GH12569).
  • Potential segfault in DataFrame.to_json when serialising datetime.time (GH11473).
  • Potential segfault in DataFrame.to_json when attempting to serialise 0d array (GH11299).
  • Segfault in to_json when attempting to serialise a DataFrame or Series with non-ndarray values; now supports serialization of category, sparse, and datetime64[ns, tz] dtypes (GH10778).
  • Bug in DataFrame.to_json with unsupported dtype not passed to default handler (GH12554).
  • Bug in .align not returning the sub-class (GH12983)
  • Bug in aligning a Series with a DataFrame (GH13037)
  • Bug in ABCPanel in which Panel4D was not being considered as a valid instance of this generic type (GH12810)
  • Bug in consistency of .name on .groupby(..).apply(..) cases (GH12363)
  • Bug in Timestamp.__repr__ that caused pprint to fail in nested structures (GH12622)
  • Bug in Timedelta.min and Timedelta.max, the properties now report the true minimum/maximum timedeltas as recognized by pandas. See the documentation. (GH12727)
  • Bug in .quantile() with interpolation may coerce to float unexpectedly (GH12772)
  • Bug in .quantile() with empty Series may return scalar rather than empty Series (GH12772)
  • Bug in .loc with out-of-bounds in a large indexer would raise IndexError rather than KeyError (GH12527)
  • Bug in resampling when using a TimedeltaIndex and .asfreq(), would previously not include the final fencepost (GH12926)
  • Bug in equality testing with a Categorical in a DataFrame (GH12564)
  • Bug in GroupBy.first(), .last() returns incorrect row when TimeGrouper is used (GH7453)
  • Bug in pd.read_csv() with the c engine when specifying skiprows with newlines in quoted items (GH10911, GH12775)
  • Bug in DataFrame timezone lost when assigning tz-aware datetime Series with alignment (GH12981)
  • Bug in .value_counts() when normalize=True and dropna=True where nulls still contributed to the normalized count (GH12558)
  • Bug in Series.value_counts() loses name if its dtype is category (GH12835)
  • Bug in Series.value_counts() loses timezone info (GH12835)
  • Bug in Series.value_counts(normalize=True) with Categorical raises UnboundLocalError (GH12835)
  • Bug in Panel.fillna() ignoring inplace=True (GH12633)
  • Bug in pd.read_csv() when specifying names, usecols, and parse_dates simultaneously with the c engine (GH9755)
  • Bug in pd.read_csv() when specifying delim_whitespace=True and lineterminator simultaneously with the c engine (GH12912)
  • Bug in Series.rename, DataFrame.rename and DataFrame.rename_axis not treating Series as mappings to relabel (GH12623).
  • Clean in .rolling.min and .rolling.max to enhance dtype handling (GH12373)
  • Bug in groupby where complex types are coerced to float (GH12902)
  • Bug in Series.map raises TypeError if its dtype is category or tz-aware datetime (GH12473)
  • Bugs on 32bit platforms for some test comparisons (GH12972)
  • Bug in index coercion when falling back from RangeIndex construction (GH12893)
  • Better error message in window functions when invalid argument (e.g. a float window) is passed (GH12669)
  • Bug in slicing subclassed DataFrame defined to return subclassed Series may return normal Series (GH11559)
  • Bug in .str accessor methods may raise ValueError if input has name and the result is DataFrame or MultiIndex (GH12617)
  • Bug in DataFrame.last_valid_index() and DataFrame.first_valid_index() on empty frames (GH12800)
  • Bug in CategoricalIndex.get_loc returns different result from regular Index (GH12531)
  • Bug in PeriodIndex.resample where name not propagated (GH12769)
  • Bug in date_range closed keyword and timezones (GH12684).
  • Bug in pd.concat raises AttributeError when input data contains tz-aware datetime and timedelta (GH12620)
  • Bug in pd.concat did not handle empty Series properly (GH11082)
  • Bug in .plot.bar alginment when width is specified with int (GH12979)
  • Bug in fill_value is ignored if the argument to a binary operator is a constant (GH12723)
  • Bug in pd.read_html() when using bs4 flavor and parsing table with a header and only one column (GH9178)
  • Bug in .pivot_table when margins=True and dropna=True where nulls still contributed to margin count (GH12577)
  • Bug in .pivot_table when dropna=False where table index/column names disappear (GH12133)
  • Bug in pd.crosstab() when margins=True and dropna=False which raised (GH12642)
  • Bug in Series.name when name attribute can be a hashable type (GH12610)
  • Bug in .describe() resets categorical columns information (GH11558)
  • Bug where loffset argument was not applied when calling resample().count() on a timeseries (GH12725)
  • pd.read_excel() now accepts column names associated with keyword argument names (GH12870)
  • Bug in pd.to_numeric() with Index returns np.ndarray, rather than Index (GH12777)
  • Bug in pd.to_numeric() with datetime-like may raise TypeError (GH12777)
  • Bug in pd.to_numeric() with scalar raises ValueError (GH12777)

v0.18.0 (March 13, 2016)

This is a major release from 0.17.1 and includes a small number of API changes, several new features, enhancements, and performance improvements along with a large number of bug fixes. We recommend that all users upgrade to this version.

Warning

pandas >= 0.18.0 no longer supports compatibility with Python version 2.6 and 3.3 (GH7718, GH11273)

Warning

numexpr version 2.4.4 will now show a warning and not be used as a computation back-end for pandas because of some buggy behavior. This does not affect other versions (>= 2.1 and >= 2.4.6). (GH12489)

Highlights include:

  • Moving and expanding window functions are now methods on Series and DataFrame, similar to .groupby, see here.
  • Adding support for a RangeIndex as a specialized form of the Int64Index for memory savings, see here.
  • API breaking change to the .resample method to make it more .groupby like, see here.
  • Removal of support for positional indexing with floats, which was deprecated since 0.14.0. This will now raise a TypeError, see here.
  • The .to_xarray() function has been added for compatibility with the xarray package, see here.
  • The read_sas function has been enhanced to read sas7bdat files, see here.
  • Addition of the .str.extractall() method, and API changes to the .str.extract() method and .str.cat() method.
  • pd.test() top-level nose test runner is available (GH4327).

Check the API Changes and deprecations before updating.

New features

Window functions are now methods

Window functions have been refactored to be methods on Series/DataFrame objects, rather than top-level functions, which are now deprecated. This allows these window-type functions, to have a similar API to that of .groupby. See the full documentation here (GH11603, GH12373)

In [1]: np.random.seed(1234)

In [2]: df = pd.DataFrame({'A' : range(10), 'B' : np.random.randn(10)})

In [3]: df
Out[3]: 
   A         B
0  0  0.471435
1  1 -1.190976
2  2  1.432707
3  3 -0.312652
4  4 -0.720589
5  5  0.887163
6  6  0.859588
7  7 -0.636524
8  8  0.015696
9  9 -2.242685

Previous Behavior:

In [8]: pd.rolling_mean(df,window=3)
        FutureWarning: pd.rolling_mean is deprecated for DataFrame and will be removed in a future version, replace with
                       DataFrame.rolling(window=3,center=False).mean()
Out[8]:
    A         B
0 NaN       NaN
1 NaN       NaN
2   1  0.237722
3   2 -0.023640
4   3  0.133155
5   4 -0.048693
6   5  0.342054
7   6  0.370076
8   7  0.079587
9   8 -0.954504

New Behavior:

In [4]: r = df.rolling(window=3)

These show a descriptive repr

In [5]: r
Out[5]: Rolling [window=3,center=False,axis=0]

with tab-completion of available methods and properties.

In [9]: r.
r.A           r.agg         r.apply       r.count       r.exclusions  r.max         r.median      r.name        r.skew        r.sum
r.B           r.aggregate   r.corr        r.cov         r.kurt        r.mean        r.min         r.quantile    r.std         r.var

The methods operate on the Rolling object itself

In [6]: r.mean()
Out[6]: 
     A         B
0  NaN       NaN
1  NaN       NaN
2  1.0  0.237722
3  2.0 -0.023640
4  3.0  0.133155
5  4.0 -0.048693
6  5.0  0.342054
7  6.0  0.370076
8  7.0  0.079587
9  8.0 -0.954504

They provide getitem accessors

In [7]: r['A'].mean()
Out[7]: 
0    NaN
1    NaN
2    1.0
3    2.0
4    3.0
5    4.0
6    5.0
7    6.0
8    7.0
9    8.0
Name: A, dtype: float64

And multiple aggregations

In [8]: r.agg({'A' : ['mean','std'],
   ...:        'B' : ['mean','std']})
   ...: 
Out[8]: 
     A              B          
  mean  std      mean       std
0  NaN  NaN       NaN       NaN
1  NaN  NaN       NaN       NaN
2  1.0  1.0  0.237722  1.327364
3  2.0  1.0 -0.023640  1.335505
4  3.0  1.0  0.133155  1.143778
5  4.0  1.0 -0.048693  0.835747
6  5.0  1.0  0.342054  0.920379
7  6.0  1.0  0.370076  0.871850
8  7.0  1.0  0.079587  0.750099
9  8.0  1.0 -0.954504  1.162285

Changes to rename

Series.rename and NDFrame.rename_axis can now take a scalar or list-like argument for altering the Series or axis name, in addition to their old behaviors of altering labels. (GH9494, GH11965)

In [9]: s = pd.Series(np.random.randn(5))

In [10]: s.rename('newname')
Out[10]: 
0    1.150036
1    0.991946
2    0.953324
3   -2.021255
4   -0.334077
Name: newname, dtype: float64
In [11]: df = pd.DataFrame(np.random.randn(5, 2))

In [12]: (df.rename_axis("indexname")
   ....:    .rename_axis("columns_name", axis="columns"))
   ....: 
Out[12]: 
columns_name         0         1
indexname                       
0             0.002118  0.405453
1             0.289092  1.321158
2            -1.546906 -0.202646
3            -0.655969  0.193421
4             0.553439  1.318152

The new functionality works well in method chains. Previously these methods only accepted functions or dicts mapping a label to a new label. This continues to work as before for function or dict-like values.

Range Index

A RangeIndex has been added to the Int64Index sub-classes to support a memory saving alternative for common use cases. This has a similar implementation to the python range object (xrange in python 2), in that it only stores the start, stop, and step values for the index. It will transparently interact with the user API, converting to Int64Index if needed.

This will now be the default constructed index for NDFrame objects, rather than previous an Int64Index. (GH939, GH12070, GH12071, GH12109, GH12888)

Previous Behavior:

In [3]: s = pd.Series(range(1000))

In [4]: s.index
Out[4]:
Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
            ...
            990, 991, 992, 993, 994, 995, 996, 997, 998, 999], dtype='int64', length=1000)

In [6]: s.index.nbytes
Out[6]: 8000

New Behavior:

In [13]: s = pd.Series(range(1000))

In [14]: s.index
Out[14]: RangeIndex(start=0, stop=1000, step=1)

In [15]: s.index.nbytes
Out[15]: 72

Changes to str.extract

The .str.extract method takes a regular expression with capture groups, finds the first match in each subject string, and returns the contents of the capture groups (GH11386).

In v0.18.0, the expand argument was added to extract.

  • expand=False: it returns a Series, Index, or DataFrame, depending on the subject and regular expression pattern (same behavior as pre-0.18.0).
  • expand=True: it always returns a DataFrame, which is more consistent and less confusing from the perspective of a user.

Currently the default is expand=None which gives a FutureWarning and uses expand=False. To avoid this warning, please explicitly specify expand.

In [1]: pd.Series(['a1', 'b2', 'c3']).str.extract('[ab](\d)', expand=None)
FutureWarning: currently extract(expand=None) means expand=False (return Index/Series/DataFrame)
but in a future version of pandas this will be changed to expand=True (return DataFrame)

Out[1]:
0      1
1      2
2    NaN
dtype: object

Extracting a regular expression with one group returns a Series if expand=False.

In [16]: pd.Series(['a1', 'b2', 'c3']).str.extract('[ab](\d)', expand=False)
Out[16]: 
0      1
1      2
2    NaN
dtype: object

It returns a DataFrame with one column if expand=True.

In [17]: pd.Series(['a1', 'b2', 'c3']).str.extract('[ab](\d)', expand=True)
Out[17]: 
     0
0    1
1    2
2  NaN

Calling on an Index with a regex with exactly one capture group returns an Index if expand=False.

In [18]: s = pd.Series(["a1", "b2", "c3"], ["A11", "B22", "C33"])

In [19]: s.index
Out[19]: Index([u'A11', u'B22', u'C33'], dtype='object')

In [20]: s.index.str.extract("(?P<letter>[a-zA-Z])", expand=False)
Out[20]: Index([u'A', u'B', u'C'], dtype='object', name=u'letter')

It returns a DataFrame with one column if expand=True.

In [21]: s.index.str.extract("(?P<letter>[a-zA-Z])", expand=True)
Out[21]: 
  letter
0      A
1      B
2      C

Calling on an Index with a regex with more than one capture group raises ValueError if expand=False.

>>> s.index.str.extract("(?P<letter>[a-zA-Z])([0-9]+)", expand=False)
ValueError: only one regex group is supported with Index

It returns a DataFrame if expand=True.

In [22]: s.index.str.extract("(?P<letter>[a-zA-Z])([0-9]+)", expand=True)
Out[22]: 
  letter   1
0      A  11
1      B  22
2      C  33

In summary, extract(expand=True) always returns a DataFrame with a row for every subject string, and a column for every capture group.

Addition of str.extractall

The .str.extractall method was added (GH11386). Unlike extract, which returns only the first match.

In [23]: s = pd.Series(["a1a2", "b1", "c1"], ["A", "B", "C"])

In [24]: s
Out[24]: 
A    a1a2
B      b1
C      c1
dtype: object

In [25]: s.str.extract("(?P<letter>[ab])(?P<digit>\d)", expand=False)
Out[25]: 
  letter digit
A      a     1
B      b     1
C    NaN   NaN

The extractall method returns all matches.

In [26]: s.str.extractall("(?P<letter>[ab])(?P<digit>\d)")
Out[26]: 
        letter digit
  match             
A 0          a     1
  1          a     2
B 0          b     1

Changes to str.cat

The method .str.cat() concatenates the members of a Series. Before, if NaN values were present in the Series, calling .str.cat() on it would return NaN, unlike the rest of the Series.str.* API. This behavior has been amended to ignore NaN values by default. (GH11435).

A new, friendlier ValueError is added to protect against the mistake of supplying the sep as an arg, rather than as a kwarg. (GH11334).

In [27]: pd.Series(['a','b',np.nan,'c']).str.cat(sep=' ')
Out[27]: 'a b c'

In [28]: pd.Series(['a','b',np.nan,'c']).str.cat(sep=' ', na_rep='?')
Out[28]: 'a b ? c'
In [2]: pd.Series(['a','b',np.nan,'c']).str.cat(' ')
ValueError: Did you mean to supply a `sep` keyword?

Datetimelike rounding

DatetimeIndex, Timestamp, TimedeltaIndex, Timedelta have gained the .round(), .floor() and .ceil() method for datetimelike rounding, flooring and ceiling. (GH4314, GH11963)

Naive datetimes

In [29]: dr = pd.date_range('20130101 09:12:56.1234', periods=3)

In [30]: dr
Out[30]: 
DatetimeIndex(['2013-01-01 09:12:56.123400', '2013-01-02 09:12:56.123400',
               '2013-01-03 09:12:56.123400'],
              dtype='datetime64[ns]', freq='D')

In [31]: dr.round('s')
Out[31]: 
DatetimeIndex(['2013-01-01 09:12:56', '2013-01-02 09:12:56',
               '2013-01-03 09:12:56'],
              dtype='datetime64[ns]', freq=None)

# Timestamp scalar
In [32]: dr[0]
Out[32]: Timestamp('2013-01-01 09:12:56.123400', freq='D')

In [33]: dr[0].round('10s')
Out[33]: Timestamp('2013-01-01 09:13:00')

Tz-aware are rounded, floored and ceiled in local times

In [34]: dr = dr.tz_localize('US/Eastern')

In [35]: dr
Out[35]: 
DatetimeIndex(['2013-01-01 09:12:56.123400-05:00',
               '2013-01-02 09:12:56.123400-05:00',
               '2013-01-03 09:12:56.123400-05:00'],
              dtype='datetime64[ns, US/Eastern]', freq='D')

In [36]: dr.round('s')
Out[36]: 
DatetimeIndex(['2013-01-01 09:12:56-05:00', '2013-01-02 09:12:56-05:00',
               '2013-01-03 09:12:56-05:00'],
              dtype='datetime64[ns, US/Eastern]', freq=None)

Timedeltas

In [37]: t = timedelta_range('1 days 2 hr 13 min 45 us',periods=3,freq='d')

In [38]: t
Out[38]: 
TimedeltaIndex(['1 days 02:13:00.000045', '2 days 02:13:00.000045',
                '3 days 02:13:00.000045'],
               dtype='timedelta64[ns]', freq='D')

In [39]: t.round('10min')
Out[39]: TimedeltaIndex(['1 days 02:10:00', '2 days 02:10:00', '3 days 02:10:00'], dtype='timedelta64[ns]', freq=None)

# Timedelta scalar
In [40]: t[0]
Out[40]: Timedelta('1 days 02:13:00.000045')

In [41]: t[0].round('2h')
Out[41]: Timedelta('1 days 02:00:00')

In addition, .round(), .floor() and .ceil() will be available thru the .dt accessor of Series.

In [42]: s = pd.Series(dr)

In [43]: s
Out[43]: 
0   2013-01-01 09:12:56.123400-05:00
1   2013-01-02 09:12:56.123400-05:00
2   2013-01-03 09:12:56.123400-05:00
dtype: datetime64[ns, US/Eastern]

In [44]: s.dt.round('D')
Out[44]: 
0   2013-01-01 00:00:00-05:00
1   2013-01-02 00:00:00-05:00
2   2013-01-03 00:00:00-05:00
dtype: datetime64[ns, US/Eastern]

Formatting of Integers in FloatIndex

Integers in FloatIndex, e.g. 1., are now formatted with a decimal point and a 0 digit, e.g. 1.0 (GH11713) This change not only affects the display to the console, but also the output of IO methods like .to_csv or .to_html.

Previous Behavior:

In [2]: s = pd.Series([1,2,3], index=np.arange(3.))

In [3]: s
Out[3]:
0    1
1    2
2    3
dtype: int64

In [4]: s.index
Out[4]: Float64Index([0.0, 1.0, 2.0], dtype='float64')

In [5]: print(s.to_csv(path=None))
0,1
1,2
2,3

New Behavior:

In [45]: s = pd.Series([1,2,3], index=np.arange(3.))

In [46]: s
Out[46]: 
0.0    1
1.0    2
2.0    3
dtype: int64

In [47]: s.index
Out[47]: Float64Index([0.0, 1.0, 2.0], dtype='float64')

In [48]: print(s.to_csv(path=None))
0.0,1
1.0,2
2.0,3

Changes to dtype assignment behaviors

When a DataFrame’s slice is updated with a new slice of the same dtype, the dtype of the DataFrame will now remain the same. (GH10503)

Previous Behavior:

In [5]: df = pd.DataFrame({'a': [0, 1, 1],
                           'b': pd.Series([100, 200, 300], dtype='uint32')})

In [7]: df.dtypes
Out[7]:
a     int64
b    uint32
dtype: object

In [8]: ix = df['a'] == 1

In [9]: df.loc[ix, 'b'] = df.loc[ix, 'b']

In [11]: df.dtypes
Out[11]:
a    int64
b    int64
dtype: object

New Behavior:

In [49]: df = pd.DataFrame({'a': [0, 1, 1],
   ....:                    'b': pd.Series([100, 200, 300], dtype='uint32')})
   ....: 

In [50]: df.dtypes
Out[50]: 
a     int64
b    uint32
dtype: object

In [51]: ix = df['a'] == 1

In [52]: df.loc[ix, 'b'] = df.loc[ix, 'b']

In [53]: df.dtypes
Out[53]: 
a     int64
b    uint32
dtype: object

When a DataFrame’s integer slice is partially updated with a new slice of floats that could potentially be downcasted to integer without losing precision, the dtype of the slice will be set to float instead of integer.

Previous Behavior:

In [4]: df = pd.DataFrame(np.array(range(1,10)).reshape(3,3),
                          columns=list('abc'),
                          index=[[4,4,8], [8,10,12]])

In [5]: df
Out[5]:
      a  b  c
4 8   1  2  3
  10  4  5  6
8 12  7  8  9

In [7]: df.ix[4, 'c'] = np.array([0., 1.])

In [8]: df
Out[8]:
      a  b  c
4 8   1  2  0
  10  4  5  1
8 12  7  8  9

New Behavior:

In [54]: df = pd.DataFrame(np.array(range(1,10)).reshape(3,3),
   ....:                   columns=list('abc'),
   ....:                   index=[[4,4,8], [8,10,12]])
   ....: 

In [55]: df
Out[55]: 
      a  b  c
4 8   1  2  3
  10  4  5  6
8 12  7  8  9

In [56]: df.ix[4, 'c'] = np.array([0., 1.])

In [57]: df
Out[57]: 
      a  b    c
4 8   1  2  0.0
  10  4  5  1.0
8 12  7  8  9.0

to_xarray

In a future version of pandas, we will be deprecating Panel and other > 2 ndim objects. In order to provide for continuity, all NDFrame objects have gained the .to_xarray() method in order to convert to xarray objects, which has a pandas-like interface for > 2 ndim. (GH11972)

See the xarray full-documentation here.

In [1]: p = Panel(np.arange(2*3*4).reshape(2,3,4))

In [2]: p.to_xarray()
Out[2]:
<xarray.DataArray (items: 2, major_axis: 3, minor_axis: 4)>
array([[[ 0,  1,  2,  3],
        [ 4,  5,  6,  7],
        [ 8,  9, 10, 11]],

       [[12, 13, 14, 15],
        [16, 17, 18, 19],
        [20, 21, 22, 23]]])
Coordinates:
  * items       (items) int64 0 1
  * major_axis  (major_axis) int64 0 1 2
  * minor_axis  (minor_axis) int64 0 1 2 3

Latex Representation

DataFrame has gained a ._repr_latex_() method in order to allow for conversion to latex in a ipython/jupyter notebook using nbconvert. (GH11778)

Note that this must be activated by setting the option pd.display.latex.repr=True (GH12182)

For example, if you have a jupyter notebook you plan to convert to latex using nbconvert, place the statement pd.display.latex.repr=True in the first cell to have the contained DataFrame output also stored as latex.

The options display.latex.escape and display.latex.longtable have also been added to the configuration and are used automatically by the to_latex method. See the available options docs for more info.

pd.read_sas() changes

read_sas has gained the ability to read SAS7BDAT files, including compressed files. The files can be read in entirety, or incrementally. For full details see here. (GH4052)

Other enhancements

  • Handle truncated floats in SAS xport files (GH11713)
  • Added option to hide index in Series.to_string (GH11729)
  • read_excel now supports s3 urls of the format s3://bucketname/filename (GH11447)
  • add support for AWS_S3_HOST env variable when reading from s3 (GH12198)
  • A simple version of Panel.round() is now implemented (GH11763)
  • For Python 3.x, round(DataFrame), round(Series), round(Panel) will work (GH11763)
  • sys.getsizeof(obj) returns the memory usage of a pandas object, including the values it contains (GH11597)
  • Series gained an is_unique attribute (GH11946)
  • DataFrame.quantile and Series.quantile now accept interpolation keyword (GH10174).
  • Added DataFrame.style.format for more flexible formatting of cell values (GH11692)
  • DataFrame.select_dtypes now allows the np.float16 typecode (GH11990)
  • pivot_table() now accepts most iterables for the values parameter (GH12017)
  • Added Google BigQuery service account authentication support, which enables authentication on remote servers. (GH11881, GH12572). For further details see here
  • HDFStore is now iterable: for k in store is equivalent to for k in store.keys() (GH12221).
  • Add missing methods/fields to .dt for Period (GH8848)
  • The entire codebase has been PEP-ified (GH12096)

Backwards incompatible API changes

  • the leading whitespaces have been removed from the output of .to_string(index=False) method (GH11833)
  • the out parameter has been removed from the Series.round() method. (GH11763)
  • DataFrame.round() leaves non-numeric columns unchanged in its return, rather than raises. (GH11885)
  • DataFrame.head(0) and DataFrame.tail(0) return empty frames, rather than self. (GH11937)
  • Series.head(0) and Series.tail(0) return empty series, rather than self. (GH11937)
  • to_msgpack and read_msgpack encoding now defaults to 'utf-8'. (GH12170)
  • the order of keyword arguments to text file parsing functions (.read_csv(), .read_table(), .read_fwf()) changed to group related arguments. (GH11555)
  • NaTType.isoformat now returns the string 'NaT to allow the result to be passed to the constructor of Timestamp. (GH12300)

NaT and Timedelta operations

NaT and Timedelta have expanded arithmetic operations, which are extended to Series arithmetic where applicable. Operations defined for datetime64[ns] or timedelta64[ns] are now also defined for NaT (GH11564).

NaT now supports arithmetic operations with integers and floats.

In [58]: pd.NaT * 1
Out[58]: NaT

In [59]: pd.NaT * 1.5
Out[59]: NaT

In [60]: pd.NaT / 2
Out[60]: NaT

In [61]: pd.NaT * np.nan
Out[61]: NaT

NaT defines more arithmetic operations with datetime64[ns] and timedelta64[ns].

In [62]: pd.NaT / pd.NaT
Out[62]: nan

In [63]: pd.Timedelta('1s') / pd.NaT
Out[63]: nan

NaT may represent either a datetime64[ns] null or a timedelta64[ns] null. Given the ambiguity, it is treated as a timedelta64[ns], which allows more operations to succeed.

In [64]: pd.NaT + pd.NaT
Out[64]: NaT

# same as
In [65]: pd.Timedelta('1s') + pd.Timedelta('1s')
Out[65]: Timedelta('0 days 00:00:02')

as opposed to

In [3]: pd.Timestamp('19900315') + pd.Timestamp('19900315')
TypeError: unsupported operand type(s) for +: 'Timestamp' and 'Timestamp'

However, when wrapped in a Series whose dtype is datetime64[ns] or timedelta64[ns], the dtype information is respected.

In [1]: pd.Series([pd.NaT], dtype='<M8[ns]') + pd.Series([pd.NaT], dtype='<M8[ns]')
TypeError: can only operate on a datetimes for subtraction,
           but the operator [__add__] was passed
In [66]: pd.Series([pd.NaT], dtype='<m8[ns]') + pd.Series([pd.NaT], dtype='<m8[ns]')
Out[66]: 
0   NaT
dtype: timedelta64[ns]

Timedelta division by floats now works.

In [67]: pd.Timedelta('1s') / 2.0
Out[67]: Timedelta('0 days 00:00:00.500000')

Subtraction by Timedelta in a Series by a Timestamp works (GH11925)

In [68]: ser = pd.Series(pd.timedelta_range('1 day', periods=3))

In [69]: ser
Out[69]: 
0   1 days
1   2 days
2   3 days
dtype: timedelta64[ns]

In [70]: pd.Timestamp('2012-01-01') - ser
Out[70]: 
0   2011-12-31
1   2011-12-30
2   2011-12-29
dtype: datetime64[ns]

NaT.isoformat() now returns 'NaT'. This change allows allows pd.Timestamp to rehydrate any timestamp like object from its isoformat (GH12300).

Changes to msgpack

Forward incompatible changes in msgpack writing format were made over 0.17.0 and 0.18.0; older versions of pandas cannot read files packed by newer versions (GH12129, GH10527)

Bugs in to_msgpack and read_msgpack introduced in 0.17.0 and fixed in 0.18.0, caused files packed in Python 2 unreadable by Python 3 (GH12142). The following table describes the backward and forward compat of msgpacks.

Warning

Packed with Can be unpacked with
pre-0.17 / Python 2 any
pre-0.17 / Python 3 any
0.17 / Python 2
  • ==0.17 / Python 2
  • >=0.18 / any Python
0.17 / Python 3 >=0.18 / any Python
0.18 >= 0.18

0.18.0 is backward-compatible for reading files packed by older versions, except for files packed with 0.17 in Python 2, in which case only they can only be unpacked in Python 2.

Signature change for .rank

Series.rank and DataFrame.rank now have the same signature (GH11759)

Previous signature

In [3]: pd.Series([0,1]).rank(method='average', na_option='keep',
                              ascending=True, pct=False)
Out[3]:
0    1
1    2
dtype: float64

In [4]: pd.DataFrame([0,1]).rank(axis=0, numeric_only=None,
                                 method='average', na_option='keep',
                                 ascending=True, pct=False)
Out[4]:
   0
0  1
1  2

New signature

In [71]: pd.Series([0,1]).rank(axis=0, method='average', numeric_only=None,
   ....:                       na_option='keep', ascending=True, pct=False)
   ....: 
Out[71]: 
0    1.0
1    2.0
dtype: float64

In [72]: pd.DataFrame([0,1]).rank(axis=0, method='average', numeric_only=None,
   ....:                          na_option='keep', ascending=True, pct=False)
   ....: 
Out[72]: 
     0
0  1.0
1  2.0

Bug in QuarterBegin with n=0

In previous versions, the behavior of the QuarterBegin offset was inconsistent depending on the date when the n parameter was 0. (GH11406)

The general semantics of anchored offsets for n=0 is to not move the date when it is an anchor point (e.g., a quarter start date), and otherwise roll forward to the next anchor point.

In [73]: d = pd.Timestamp('2014-02-01')

In [74]: d
Out[74]: Timestamp('2014-02-01 00:00:00')

In [75]: d + pd.offsets.QuarterBegin(n=0, startingMonth=2)
Out[75]: Timestamp('2014-02-01 00:00:00')

In [76]: d + pd.offsets.QuarterBegin(n=0, startingMonth=1)
Out[76]: Timestamp('2014-04-01 00:00:00')

For the QuarterBegin offset in previous versions, the date would be rolled backwards if date was in the same month as the quarter start date.

In [3]: d = pd.Timestamp('2014-02-15')

In [4]: d + pd.offsets.QuarterBegin(n=0, startingMonth=2)
Out[4]: Timestamp('2014-02-01 00:00:00')

This behavior has been corrected in version 0.18.0, which is consistent with other anchored offsets like MonthBegin and YearBegin.

In [77]: d = pd.Timestamp('2014-02-15')

In [78]: d + pd.offsets.QuarterBegin(n=0, startingMonth=2)
Out[78]: Timestamp('2014-05-01 00:00:00')

Resample API

Like the change in the window functions API above, .resample(...) is changing to have a more groupby-like API. (GH11732, GH12702, GH12202, GH12332, GH12334, GH12348, GH12448).

In [79]: np.random.seed(1234)

In [80]: df = pd.DataFrame(np.random.rand(10,4),
   ....:                   columns=list('ABCD'),
   ....:                   index=pd.date_range('2010-01-01 09:00:00', periods=10, freq='s'))
   ....: 

In [81]: df
Out[81]: 
                            A         B         C         D
2010-01-01 09:00:00  0.191519  0.622109  0.437728  0.785359
2010-01-01 09:00:01  0.779976  0.272593  0.276464  0.801872
2010-01-01 09:00:02  0.958139  0.875933  0.357817  0.500995
2010-01-01 09:00:03  0.683463  0.712702  0.370251  0.561196
2010-01-01 09:00:04  0.503083  0.013768  0.772827  0.882641
2010-01-01 09:00:05  0.364886  0.615396  0.075381  0.368824
2010-01-01 09:00:06  0.933140  0.651378  0.397203  0.788730
2010-01-01 09:00:07  0.316836  0.568099  0.869127  0.436173
2010-01-01 09:00:08  0.802148  0.143767  0.704261  0.704581
2010-01-01 09:00:09  0.218792  0.924868  0.442141  0.909316

Previous API:

You would write a resampling operation that immediately evaluates. If a how parameter was not provided, it would default to how='mean'.

In [6]: df.resample('2s')
Out[6]:
                         A         B         C         D
2010-01-01 09:00:00  0.485748  0.447351  0.357096  0.793615
2010-01-01 09:00:02  0.820801  0.794317  0.364034  0.531096
2010-01-01 09:00:04  0.433985  0.314582  0.424104  0.625733
2010-01-01 09:00:06  0.624988  0.609738  0.633165  0.612452
2010-01-01 09:00:08  0.510470  0.534317  0.573201  0.806949

You could also specify a how directly

In [7]: df.resample('2s', how='sum')
Out[7]:
                         A         B         C         D
2010-01-01 09:00:00  0.971495  0.894701  0.714192  1.587231
2010-01-01 09:00:02  1.641602  1.588635  0.728068  1.062191
2010-01-01 09:00:04  0.867969  0.629165  0.848208  1.251465
2010-01-01 09:00:06  1.249976  1.219477  1.266330  1.224904
2010-01-01 09:00:08  1.020940  1.068634  1.146402  1.613897

New API:

Now, you can write .resample(..) as a 2-stage operation like .groupby(...), which yields a Resampler.

In [82]: r = df.resample('2s')

In [83]: r
Out[83]: DatetimeIndexResampler [freq=<2 * Seconds>, axis=0, closed=left, label=left, convention=start, base=0]
Downsampling

You can then use this object to perform operations. These are downsampling operations (going from a higher frequency to a lower one).

In [84]: r.mean()
Out[84]: 
                            A         B         C         D
2010-01-01 09:00:00  0.485748  0.447351  0.357096  0.793615
2010-01-01 09:00:02  0.820801  0.794317  0.364034  0.531096
2010-01-01 09:00:04  0.433985  0.314582  0.424104  0.625733
2010-01-01 09:00:06  0.624988  0.609738  0.633165  0.612452
2010-01-01 09:00:08  0.510470  0.534317  0.573201  0.806949
In [85]: r.sum()
Out[85]: 
                            A         B         C         D
2010-01-01 09:00:00  0.971495  0.894701  0.714192  1.587231
2010-01-01 09:00:02  1.641602  1.588635  0.728068  1.062191
2010-01-01 09:00:04  0.867969  0.629165  0.848208  1.251465
2010-01-01 09:00:06  1.249976  1.219477  1.266330  1.224904
2010-01-01 09:00:08  1.020940  1.068634  1.146402  1.613897

Furthermore, resample now supports getitem operations to perform the resample on specific columns.

In [86]: r[['A','C']].mean()
Out[86]: 
                            A         C
2010-01-01 09:00:00  0.485748  0.357096
2010-01-01 09:00:02  0.820801  0.364034
2010-01-01 09:00:04  0.433985  0.424104
2010-01-01 09:00:06  0.624988  0.633165
2010-01-01 09:00:08  0.510470  0.573201

and .aggregate type operations.

In [87]: r.agg({'A' : 'mean', 'B' : 'sum'})
Out[87]: 
                            A         B
2010-01-01 09:00:00  0.485748  0.894701
2010-01-01 09:00:02  0.820801  1.588635
2010-01-01 09:00:04  0.433985  0.629165
2010-01-01 09:00:06  0.624988  1.219477
2010-01-01 09:00:08  0.510470  1.068634

These accessors can of course, be combined

In [88]: r[['A','B']].agg(['mean','sum'])
Out[88]: 
                            A                   B          
                         mean       sum      mean       sum
2010-01-01 09:00:00  0.485748  0.971495  0.447351  0.894701
2010-01-01 09:00:02  0.820801  1.641602  0.794317  1.588635
2010-01-01 09:00:04  0.433985  0.867969  0.314582  0.629165
2010-01-01 09:00:06  0.624988  1.249976  0.609738  1.219477
2010-01-01 09:00:08  0.510470  1.020940  0.534317  1.068634
Upsampling

Upsampling operations take you from a lower frequency to a higher frequency. These are now performed with the Resampler objects with backfill(), ffill(), fillna() and asfreq() methods.

In [89]: s = pd.Series(np.arange(5,dtype='int64'),
   ....:               index=date_range('2010-01-01', periods=5, freq='Q'))
   ....: 

In [90]: s
Out[90]: 
2010-03-31    0
2010-06-30    1
2010-09-30    2
2010-12-31    3
2011-03-31    4
Freq: Q-DEC, dtype: int64

Previously

In [6]: s.resample('M', fill_method='ffill')
Out[6]:
2010-03-31    0
2010-04-30    0
2010-05-31    0
2010-06-30    1
2010-07-31    1
2010-08-31    1
2010-09-30    2
2010-10-31    2
2010-11-30    2
2010-12-31    3
2011-01-31    3
2011-02-28    3
2011-03-31    4
Freq: M, dtype: int64

New API

In [91]: s.resample('M').ffill()
Out[91]: 
2010-03-31    0
2010-04-30    0
2010-05-31    0
2010-06-30    1
2010-07-31    1
2010-08-31    1
2010-09-30    2
2010-10-31    2
2010-11-30    2
2010-12-31    3
2011-01-31    3
2011-02-28    3
2011-03-31    4
Freq: M, dtype: int64

Note

In the new API, you can either downsample OR upsample. The prior implementation would allow you to pass an aggregator function (like mean) even though you were upsampling, providing a bit of confusion.

Previous API will work but with deprecations

Warning

This new API for resample includes some internal changes for the prior-to-0.18.0 API, to work with a deprecation warning in most cases, as the resample operation returns a deferred object. We can intercept operations and just do what the (pre 0.18.0) API did (with a warning). Here is a typical use case:

In [4]: r = df.resample('2s')

In [6]: r*10
pandas/tseries/resample.py:80: FutureWarning: .resample() is now a deferred operation
use .resample(...).mean() instead of .resample(...)

Out[6]:
                      A         B         C         D
2010-01-01 09:00:00  4.857476  4.473507  3.570960  7.936154
2010-01-01 09:00:02  8.208011  7.943173  3.640340  5.310957
2010-01-01 09:00:04  4.339846  3.145823  4.241039  6.257326
2010-01-01 09:00:06  6.249881  6.097384  6.331650  6.124518
2010-01-01 09:00:08  5.104699  5.343172  5.732009  8.069486

However, getting and assignment operations directly on a Resampler will raise a ValueError:

In [7]: r.iloc[0] = 5
ValueError: .resample() is now a deferred operation
use .resample(...).mean() instead of .resample(...)

There is a situation where the new API can not perform all the operations when using original code. This code is intending to resample every 2s, take the mean AND then take the min of those results.

In [4]: df.resample('2s').min()
Out[4]:
A    0.433985
B    0.314582
C    0.357096
D    0.531096
dtype: float64

The new API will:

In [92]: df.resample('2s').min()
Out[92]: 
                            A         B         C         D
2010-01-01 09:00:00  0.191519  0.272593  0.276464  0.785359
2010-01-01 09:00:02  0.683463  0.712702  0.357817  0.500995
2010-01-01 09:00:04  0.364886  0.013768  0.075381  0.368824
2010-01-01 09:00:06  0.316836  0.568099  0.397203  0.436173
2010-01-01 09:00:08  0.218792  0.143767  0.442141  0.704581

The good news is the return dimensions will differ between the new API and the old API, so this should loudly raise an exception.

To replicate the original operation

In [93]: df.resample('2s').mean().min()
Out[93]: 
A    0.433985
B    0.314582
C    0.357096
D    0.531096
dtype: float64

Changes to eval

In prior versions, new columns assignments in an eval expression resulted in an inplace change to the DataFrame. (GH9297, GH8664, GH10486)

In [94]: df = pd.DataFrame({'a': np.linspace(0, 10, 5), 'b': range(5)})

In [95]: df
Out[95]: 
      a  b
0   0.0  0
1   2.5  1
2   5.0  2
3   7.5  3
4  10.0  4
In [12]: df.eval('c = a + b')
FutureWarning: eval expressions containing an assignment currentlydefault to operating inplace.
This will change in a future version of pandas, use inplace=True to avoid this warning.

In [13]: df
Out[13]:
      a  b     c
0   0.0  0   0.0
1   2.5  1   3.5
2   5.0  2   7.0
3   7.5  3  10.5
4  10.0  4  14.0

In version 0.18.0, a new inplace keyword was added to choose whether the assignment should be done inplace or return a copy.

In [96]: df
Out[96]: 
      a  b     c
0   0.0  0   0.0
1   2.5  1   3.5
2   5.0  2   7.0
3   7.5  3  10.5
4  10.0  4  14.0

In [97]: df.eval('d = c - b', inplace=False)
Out[97]: 
      a  b     c     d
0   0.0  0   0.0   0.0
1   2.5  1   3.5   2.5
2   5.0  2   7.0   5.0
3   7.5  3  10.5   7.5
4  10.0  4  14.0  10.0

In [98]: df
Out[98]: 
      a  b     c
0   0.0  0   0.0
1   2.5  1   3.5
2   5.0  2   7.0
3   7.5  3  10.5
4  10.0  4  14.0

In [99]: df.eval('d = c - b', inplace=True)

In [100]: df
Out[100]: 
      a  b     c     d
0   0.0  0   0.0   0.0
1   2.5  1   3.5   2.5
2   5.0  2   7.0   5.0
3   7.5  3  10.5   7.5
4  10.0  4  14.0  10.0

Warning

For backwards compatability, inplace defaults to True if not specified. This will change in a future version of pandas. If your code depends on an inplace assignment you should update to explicitly set inplace=True

The inplace keyword parameter was also added the query method.

In [101]: df.query('a > 5')
Out[101]: 
      a  b     c     d
3   7.5  3  10.5   7.5
4  10.0  4  14.0  10.0

In [102]: df.query('a > 5', inplace=True)

In [103]: df
Out[103]: 
      a  b     c     d
3   7.5  3  10.5   7.5
4  10.0  4  14.0  10.0

Warning

Note that the default value for inplace in a query is False, which is consistent with prior versions.

eval has also been updated to allow multi-line expressions for multiple assignments. These expressions will be evaluated one at a time in order. Only assignments are valid for multi-line expressions.

In [104]: df
Out[104]: 
      a  b     c     d
3   7.5  3  10.5   7.5
4  10.0  4  14.0  10.0

In [105]: df.eval("""
   .....: e = d + a
   .....: f = e - 22
   .....: g = f / 2.0""", inplace=True)
   .....: 

In [106]: df
Out[106]: 
      a  b     c     d     e    f    g
3   7.5  3  10.5   7.5  15.0 -7.0 -3.5
4  10.0  4  14.0  10.0  20.0 -2.0 -1.0

Other API Changes

  • DataFrame.between_time and Series.between_time now only parse a fixed set of time strings. Parsing of date strings is no longer supported and raises a ValueError. (GH11818)

    In [107]: s = pd.Series(range(10), pd.date_range('2015-01-01', freq='H', periods=10))
    
    In [108]: s.between_time("7:00am", "9:00am")
    ---------------------------------------------------------------------------
    ValueError                                Traceback (most recent call last)
    <ipython-input-108-1f395af72989> in <module>()
    ----> 1 s.between_time("7:00am", "9:00am")
    
    /home/joris/scipy/pandas/pandas/core/generic.pyc in between_time(self, start_time, end_time, include_start, include_end)
       4054             indexer = self.index.indexer_between_time(
       4055                 start_time, end_time, include_start=include_start,
    -> 4056                 include_end=include_end)
       4057             return self.take(indexer, convert=False)
       4058         except AttributeError:
    
    /home/joris/scipy/pandas/pandas/tseries/index.pyc in indexer_between_time(self, start_time, end_time, include_start, include_end)
       1879         values_between_time : TimeSeries
       1880         """
    -> 1881         start_time = to_time(start_time)
       1882         end_time = to_time(end_time)
       1883         time_micros = self._get_time_micros()
    
    /home/joris/scipy/pandas/pandas/tseries/tools.pyc in to_time(arg, format, infer_time_format, errors)
        766         return _convert_listlike(arg, format)
        767 
    --> 768     return _convert_listlike(np.array([arg]), format)[0]
        769 
        770 
    
    /home/joris/scipy/pandas/pandas/tseries/tools.pyc in _convert_listlike(arg, format)
        746                 elif errors == 'raise':
        747                     raise ValueError("Cannot convert arg {arg} to "
    --> 748                                      "a time".format(arg=arg))
        749                 elif errors == 'ignore':
        750                     return arg
    
    ValueError: Cannot convert arg ['7:00am'] to a time
    

    This will now raise.

    In [2]: s.between_time('20150101 07:00:00','20150101 09:00:00')
    ValueError: Cannot convert arg ['20150101 07:00:00'] to a time.
    
  • .memory_usage() now includes values in the index, as does memory_usage in .info() (GH11597)

  • DataFrame.to_latex() now supports non-ascii encodings (eg utf-8) in Python 2 with the parameter encoding (GH7061)

  • pandas.merge() and DataFrame.merge() will show a specific error message when trying to merge with an object that is not of type DataFrame or a subclass (GH12081)

  • DataFrame.unstack and Series.unstack now take fill_value keyword to allow direct replacement of missing values when an unstack results in missing values in the resulting DataFrame. As an added benefit, specifying fill_value will preserve the data type of the original stacked data. (GH9746)

  • As part of the new API for window functions and resampling, aggregation functions have been clarified, raising more informative error messages on invalid aggregations. (GH9052). A full set of examples are presented in groupby.

  • Statistical functions for NDFrame objects (like sum(), mean(), min()) will now raise if non-numpy-compatible arguments are passed in for **kwargs (GH12301)

  • .to_latex and .to_html gain a decimal parameter like .to_csv; the default is '.' (GH12031)

  • More helpful error message when constructing a DataFrame with empty data but with indices (GH8020)

  • .describe() will now properly handle bool dtype as a categorical (GH6625)

  • More helpful error message with an invalid .transform with user defined input (GH10165)

  • Exponentially weighted functions now allow specifying alpha directly (GH10789) and raise ValueError if parameters violate 0 < alpha <= 1 (GH12492)

Deprecations

  • The functions pd.rolling_*, pd.expanding_*, and pd.ewm* are deprecated and replaced by the corresponding method call. Note that the new suggested syntax includes all of the arguments (even if default) (GH11603)

    In [1]: s = pd.Series(range(3))
    
    In [2]: pd.rolling_mean(s,window=2,min_periods=1)
            FutureWarning: pd.rolling_mean is deprecated for Series and
                 will be removed in a future version, replace with
                 Series.rolling(min_periods=1,window=2,center=False).mean()
    Out[2]:
            0    0.0
            1    0.5
            2    1.5
            dtype: float64
    
    In [3]: pd.rolling_cov(s, s, window=2)
            FutureWarning: pd.rolling_cov is deprecated for Series and
                 will be removed in a future version, replace with
                 Series.rolling(window=2).cov(other=<Series>)
    Out[3]:
            0    NaN
            1    0.5
            2    0.5
            dtype: float64
    
  • The the freq and how arguments to the .rolling, .expanding, and .ewm (new) functions are deprecated, and will be removed in a future version. You can simply resample the input prior to creating a window function. (GH11603).

    For example, instead of s.rolling(window=5,freq='D').max() to get the max value on a rolling 5 Day window, one could use s.resample('D').mean().rolling(window=5).max(), which first resamples the data to daily data, then provides a rolling 5 day window.

  • pd.tseries.frequencies.get_offset_name function is deprecated. Use offset’s .freqstr property as alternative (GH11192)

  • pandas.stats.fama_macbeth routines are deprecated and will be removed in a future version (GH6077)

  • pandas.stats.ols, pandas.stats.plm and pandas.stats.var routines are deprecated and will be removed in a future version (GH6077)

  • show a FutureWarning rather than a DeprecationWarning on using long-time deprecated syntax in HDFStore.select, where the where clause is not a string-like (GH12027)

  • The pandas.options.display.mpl_style configuration has been deprecated and will be removed in a future version of pandas. This functionality is better handled by matplotlib’s style sheets (GH11783).

Removal of deprecated float indexers

In GH4892 indexing with floating point numbers on a non-Float64Index was deprecated (in version 0.14.0). In 0.18.0, this deprecation warning is removed and these will now raise a TypeError. (GH12165, GH12333)

In [109]: s = pd.Series([1, 2, 3], index=[4, 5, 6])

In [110]: s
Out[110]: 
4    1
5    2
6    3
dtype: int64

In [111]: s2 = pd.Series([1, 2, 3], index=list('abc'))

In [112]: s2
Out[112]: 
a    1
b    2
c    3
dtype: int64

Previous Behavior:

# this is label indexing
In [2]: s[5.0]
FutureWarning: scalar indexers for index type Int64Index should be integers and not floating point
Out[2]: 2

# this is positional indexing
In [3]: s.iloc[1.0]
FutureWarning: scalar indexers for index type Int64Index should be integers and not floating point
Out[3]: 2

# this is label indexing
In [4]: s.loc[5.0]
FutureWarning: scalar indexers for index type Int64Index should be integers and not floating point
Out[4]: 2

# .ix would coerce 1.0 to the positional 1, and index
In [5]: s2.ix[1.0] = 10
FutureWarning: scalar indexers for index type Index should be integers and not floating point

In [6]: s2
Out[6]:
a     1
b    10
c     3
dtype: int64

New Behavior:

For iloc, getting & setting via a float scalar will always raise.

In [3]: s.iloc[2.0]
TypeError: cannot do label indexing on <class 'pandas.indexes.numeric.Int64Index'> with these indexers [2.0] of <type 'float'>

Other indexers will coerce to a like integer for both getting and setting. The FutureWarning has been dropped for .loc, .ix and [].

In [113]: s[5.0]
Out[113]: 2

In [114]: s.loc[5.0]
Out[114]: 2

In [115]: s.ix[5.0]
Out[115]: 2

and setting

In [116]: s_copy = s.copy()

In [117]: s_copy[5.0] = 10

In [118]: s_copy
Out[118]: 
4     1
5    10
6     3
dtype: int64

In [119]: s_copy = s.copy()

In [120]: s_copy.loc[5.0] = 10

In [121]: s_copy
Out[121]: 
4     1
5    10
6     3
dtype: int64

In [122]: s_copy = s.copy()

In [123]: s_copy.ix[5.0] = 10

In [124]: s_copy
Out[124]: 
4     1
5    10
6     3
dtype: int64

Positional setting with .ix and a float indexer will ADD this value to the index, rather than previously setting the value by position.

In [125]: s2.ix[1.0] = 10

In [126]: s2
Out[126]: 
a       1
b       2
c       3
1.0    10
dtype: int64

Slicing will also coerce integer-like floats to integers for a non-Float64Index.

In [127]: s.loc[5.0:6]
Out[127]: 
5    2
6    3
dtype: int64

In [128]: s.ix[5.0:6]
Out[128]: 
5    2
6    3
dtype: int64

Note that for floats that are NOT coercible to ints, the label based bounds will be excluded

In [129]: s.loc[5.1:6]
Out[129]: 
6    3
dtype: int64

In [130]: s.ix[5.1:6]
Out[130]: 
6    3
dtype: int64

Float indexing on a Float64Index is unchanged.

In [131]: s = pd.Series([1, 2, 3], index=np.arange(3.))

In [132]: s[1.0]
Out[132]: 2

In [133]: s[1.0:2.5]
Out[133]: 
1.0    2
2.0    3
dtype: int64

Removal of prior version deprecations/changes

  • Removal of rolling_corr_pairwise in favor of .rolling().corr(pairwise=True) (GH4950)
  • Removal of expanding_corr_pairwise in favor of .expanding().corr(pairwise=True) (GH4950)
  • Removal of DataMatrix module. This was not imported into the pandas namespace in any event (GH12111)
  • Removal of cols keyword in favor of subset in DataFrame.duplicated() and DataFrame.drop_duplicates() (GH6680)
  • Removal of the read_frame and frame_query (both aliases for pd.read_sql) and write_frame (alias of to_sql) functions in the pd.io.sql namespace, deprecated since 0.14.0 (GH6292).
  • Removal of the order keyword from .factorize() (GH6930)

Performance Improvements

  • Improved performance of andrews_curves (GH11534)
  • Improved huge DatetimeIndex, PeriodIndex and TimedeltaIndex‘s ops performance including NaT (GH10277)
  • Improved performance of pandas.concat (GH11958)
  • Improved performance of StataReader (GH11591)
  • Improved performance in construction of Categoricals with Series of datetimes containing NaT (GH12077)
  • Improved performance of ISO 8601 date parsing for dates without separators (GH11899), leading zeros (GH11871) and with whitespace preceding the time zone (GH9714)

Bug Fixes

  • Bug in GroupBy.size when data-frame is empty. (GH11699)
  • Bug in Period.end_time when a multiple of time period is requested (GH11738)
  • Regression in .clip with tz-aware datetimes (GH11838)
  • Bug in date_range when the boundaries fell on the frequency (GH11804, GH12409)
  • Bug in consistency of passing nested dicts to .groupby(...).agg(...) (GH9052)
  • Accept unicode in Timedelta constructor (GH11995)
  • Bug in value label reading for StataReader when reading incrementally (GH12014)
  • Bug in vectorized DateOffset when n parameter is 0 (GH11370)
  • Compat for numpy 1.11 w.r.t. NaT comparison changes (GH12049)
  • Bug in read_csv when reading from a StringIO in threads (GH11790)
  • Bug in not treating NaT as a missing value in datetimelikes when factorizing & with Categoricals (GH12077)
  • Bug in getitem when the values of a Series were tz-aware (GH12089)
  • Bug in Series.str.get_dummies when one of the variables was ‘name’ (GH12180)
  • Bug in pd.concat while concatenating tz-aware NaT series. (GH11693, GH11755, GH12217)
  • Bug in pd.read_stata with version <= 108 files (GH12232)
  • Bug in Series.resample using a frequency of Nano when the index is a DatetimeIndex and contains non-zero nanosecond parts (GH12037)
  • Bug in resampling with .nunique and a sparse index (GH12352)
  • Removed some compiler warnings (GH12471)
  • Work around compat issues with boto in python 3.5 (GH11915)
  • Bug in NaT subtraction from Timestamp or DatetimeIndex with timezones (GH11718)
  • Bug in subtraction of Series of a single tz-aware Timestamp (GH12290)
  • Use compat iterators in PY2 to support .next() (GH12299)
  • Bug in Timedelta.round with negative values (GH11690)
  • Bug in .loc against CategoricalIndex may result in normal Index (GH11586)
  • Bug in DataFrame.info when duplicated column names exist (GH11761)
  • Bug in .copy of datetime tz-aware objects (GH11794)
  • Bug in Series.apply and Series.map where timedelta64 was not boxed (GH11349)
  • Bug in DataFrame.set_index() with tz-aware Series (GH12358)
  • Bug in subclasses of DataFrame where AttributeError did not propagate (GH11808)
  • Bug groupby on tz-aware data where selection not returning Timestamp (GH11616)
  • Bug in pd.read_clipboard and pd.to_clipboard functions not supporting Unicode; upgrade included pyperclip to v1.5.15 (GH9263)
  • Bug in DataFrame.query containing an assignment (GH8664)
  • Bug in from_msgpack where __contains__() fails for columns of the unpacked DataFrame, if the DataFrame has object columns. (GH11880)
  • Bug in .resample on categorical data with TimedeltaIndex (GH12169)
  • Bug in timezone info lost when broadcasting scalar datetime to DataFrame (GH11682)
  • Bug in Index creation from Timestamp with mixed tz coerces to UTC (GH11488)
  • Bug in to_numeric where it does not raise if input is more than one dimension (GH11776)
  • Bug in parsing timezone offset strings with non-zero minutes (GH11708)
  • Bug in df.plot using incorrect colors for bar plots under matplotlib 1.5+ (GH11614)
  • Bug in the groupby plot method when using keyword arguments (GH11805).
  • Bug in DataFrame.duplicated and drop_duplicates causing spurious matches when setting keep=False (GH11864)
  • Bug in .loc result with duplicated key may have Index with incorrect dtype (GH11497)
  • Bug in pd.rolling_median where memory allocation failed even with sufficient memory (GH11696)
  • Bug in DataFrame.style with spurious zeros (GH12134)
  • Bug in DataFrame.style with integer columns not starting at 0 (GH12125)
  • Bug in .style.bar may not rendered properly using specific browser (GH11678)
  • Bug in rich comparison of Timedelta with a numpy.array of Timedelta that caused an infinite recursion (GH11835)
  • Bug in DataFrame.round dropping column index name (GH11986)
  • Bug in df.replace while replacing value in mixed dtype Dataframe (GH11698)
  • Bug in Index prevents copying name of passed Index, when a new name is not provided (GH11193)
  • Bug in read_excel failing to read any non-empty sheets when empty sheets exist and sheetname=None (GH11711)
  • Bug in read_excel failing to raise NotImplemented error when keywords parse_dates and date_parser are provided (GH11544)
  • Bug in read_sql with pymysql connections failing to return chunked data (GH11522)
  • Bug in .to_csv ignoring formatting parameters decimal, na_rep, float_format for float indexes (GH11553)
  • Bug in Int64Index and Float64Index preventing the use of the modulo operator (GH9244)
  • Bug in MultiIndex.drop for not lexsorted multi-indexes (GH12078)
  • Bug in DataFrame when masking an empty DataFrame (GH11859)
  • Bug in .plot potentially modifying the colors input when the number of columns didn’t match the number of series provided (GH12039).
  • Bug in Series.plot failing when index has a CustomBusinessDay frequency (GH7222).
  • Bug in .to_sql for datetime.time values with sqlite fallback (GH8341)
  • Bug in read_excel failing to read data with one column when squeeze=True (GH12157)
  • Bug in read_excel failing to read one empty column (GH12292, GH9002)
  • Bug in .groupby where a KeyError was not raised for a wrong column if there was only one row in the dataframe (GH11741)
  • Bug in .read_csv with dtype specified on empty data producing an error (GH12048)
  • Bug in .read_csv where strings like '2E' are treated as valid floats (GH12237)
  • Bug in building pandas with debugging symbols (GH12123)
  • Removed millisecond property of DatetimeIndex. This would always raise a ValueError (GH12019).
  • Bug in Series constructor with read-only data (GH11502)
  • Removed pandas.util.testing.choice(). Should use np.random.choice(), instead. (GH12386)
  • Bug in .loc setitem indexer preventing the use of a TZ-aware DatetimeIndex (GH12050)
  • Bug in .style indexes and multi-indexes not appearing (GH11655)
  • Bug in to_msgpack and from_msgpack which did not correctly serialize or deserialize NaT (GH12307).
  • Bug in .skew and .kurt due to roundoff error for highly similar values (GH11974)
  • Bug in Timestamp constructor where microsecond resolution was lost if HHMMSS were not separated with ‘:’ (GH10041)
  • Bug in buffer_rd_bytes src->buffer could be freed more than once if reading failed, causing a segfault (GH12098)
  • Bug in crosstab where arguments with non-overlapping indexes would return a KeyError (GH10291)
  • Bug in DataFrame.apply in which reduction was not being prevented for cases in which dtype was not a numpy dtype (GH12244)
  • Bug when initializing categorical series with a scalar value. (GH12336)
  • Bug when specifying a UTC DatetimeIndex by setting utc=True in .to_datetime (GH11934)
  • Bug when increasing the buffer size of CSV reader in read_csv (GH12494)
  • Bug when setting columns of a DataFrame with duplicate column names (GH12344)

v0.17.1 (November 21, 2015)

Note

We are proud to announce that pandas has become a sponsored project of the (NUMFocus organization). This will help ensure the success of development of pandas as a world-class open-source project.

This is a minor bug-fix release from 0.17.0 and includes a large number of bug fixes along several new features, enhancements, and performance improvements. We recommend that all users upgrade to this version.

Highlights include:

  • Support for Conditional HTML Formatting, see here
  • Releasing the GIL on the csv reader & other ops, see here
  • Fixed regression in DataFrame.drop_duplicates from 0.16.2, causing incorrect results on integer values (GH11376)

New features

Conditional HTML Formatting

Warning

This is a new feature and is under active development. We’ll be adding features an possibly making breaking changes in future releases. Feedback is welcome.

We’ve added experimental support for conditional HTML formatting: the visual styling of a DataFrame based on the data. The styling is accomplished with HTML and CSS. Acesses the styler class with the pandas.DataFrame.style, attribute, an instance of Styler with your data attached.

Here’s a quick example:

In [1]: np.random.seed(123)

In [2]: df = DataFrame(np.random.randn(10, 5), columns=list('abcde'))

In [3]: html = df.style.background_gradient(cmap='viridis', low=.5)

We can render the HTML to get the following table.

a b c d e
0 -1.085631 0.997345 0.282978 -1.506295 -0.5786
1 1.651437 -2.426679 -0.428913 1.265936 -0.86674
2 -0.678886 -0.094709 1.49139 -0.638902 -0.443982
3 -0.434351 2.20593 2.186786 1.004054 0.386186
4 0.737369 1.490732 -0.935834 1.175829 -1.253881
5 -0.637752 0.907105 -1.428681 -0.140069 -0.861755
6 -0.255619 -2.798589 -1.771533 -0.699877 0.927462
7 -0.173636 0.002846 0.688223 -0.879536 0.283627
8 -0.805367 -1.727669 -0.3909 0.573806 0.338589
9 -0.01183 2.392365 0.412912 0.978736 2.238143

Styler interacts nicely with the Jupyter Notebook. See the documentation for more.

Enhancements

  • DatetimeIndex now supports conversion to strings with astype(str) (GH10442)

  • Support for compression (gzip/bz2) in pandas.DataFrame.to_csv() (GH7615)

  • pd.read_* functions can now also accept pathlib.Path, or py._path.local.LocalPath objects for the filepath_or_buffer argument. (GH11033) - The DataFrame and Series functions .to_csv(), .to_html() and .to_latex() can now handle paths beginning with tildes (e.g. ~/Documents/) (GH11438)

  • DataFrame now uses the fields of a namedtuple as columns, if columns are not supplied (GH11181)

  • DataFrame.itertuples() now returns namedtuple objects, when possible. (GH11269, GH11625)

  • Added axvlines_kwds to parallel coordinates plot (GH10709)

  • Option to .info() and .memory_usage() to provide for deep introspection of memory consumption. Note that this can be expensive to compute and therefore is an optional parameter. (GH11595)

    In [4]: df = DataFrame({'A' : ['foo']*1000})
    
    In [5]: df['B'] = df['A'].astype('category')
    
    # shows the '+' as we have object dtypes
    In [6]: df.info()
    <class 'pandas.core.frame.DataFrame'>
    RangeIndex: 1000 entries, 0 to 999
    Data columns (total 2 columns):
    A    1000 non-null object
    B    1000 non-null category
    dtypes: category(1), object(1)
    memory usage: 8.9+ KB
    
    # we have an accurate memory assessment (but can be expensive to compute this)
    In [7]: df.info(memory_usage='deep')
    <class 'pandas.core.frame.DataFrame'>
    RangeIndex: 1000 entries, 0 to 999
    Data columns (total 2 columns):
    A    1000 non-null object
    B    1000 non-null category
    dtypes: category(1), object(1)
    memory usage: 48.0 KB
    
  • Index now has a fillna method (GH10089)

    In [8]: pd.Index([1, np.nan, 3]).fillna(2)
    Out[8]: Float64Index([1.0, 2.0, 3.0], dtype='float64')
    
  • Series of type category now make .str.<...> and .dt.<...> accessor methods / properties available, if the categories are of that type. (GH10661)

    In [9]: s = pd.Series(list('aabb')).astype('category')
    
    In [10]: s
    Out[10]: 
    0    a
    1    a
    2    b
    3    b
    dtype: category
    Categories (2, object): [a, b]
    
    In [11]: s.str.contains("a")
    Out[11]: 
    0     True
    1     True
    2    False
    3    False
    dtype: bool
    
    In [12]: date = pd.Series(pd.date_range('1/1/2015', periods=5)).astype('category')
    
    In [13]: date
    Out[13]: 
    0   2015-01-01
    1   2015-01-02
    2   2015-01-03
    3   2015-01-04
    4   2015-01-05
    dtype: category
    Categories (5, datetime64[ns]): [2015-01-01, 2015-01-02, 2015-01-03, 2015-01-04, 2015-01-05]
    
    In [14]: date.dt.day
    Out[14]: 
    0    1
    1    2
    2    3
    3    4
    4    5
    dtype: int64
    
  • pivot_table now has a margins_name argument so you can use something other than the default of ‘All’ (GH3335)

  • Implement export of datetime64[ns, tz] dtypes with a fixed HDF5 store (GH11411)

  • Pretty printing sets (e.g. in DataFrame cells) now uses set literal syntax ({x, y}) instead of Legacy Python syntax (set([x, y])) (GH11215)

  • Improve the error message in pandas.io.gbq.to_gbq() when a streaming insert fails (GH11285) and when the DataFrame does not match the schema of the destination table (GH11359)

API changes

  • raise NotImplementedError in Index.shift for non-supported index types (GH8038)
  • min and max reductions on datetime64 and timedelta64 dtyped series now result in NaT and not nan (GH11245).
  • Indexing with a null key will raise a TypeError, instead of a ValueError (GH11356)
  • Series.ptp will now ignore missing values by default (GH11163)

Deprecations

  • The pandas.io.ga module which implements google-analytics support is deprecated and will be removed in a future version (GH11308)
  • Deprecate the engine keyword in .to_csv(), which will be removed in a future version (GH11274)

Performance Improvements

  • Checking monotonic-ness before sorting on an index (GH11080)
  • Series.dropna performance improvement when its dtype can’t contain NaN (GH11159)
  • Release the GIL on most datetime field operations (e.g. DatetimeIndex.year, Series.dt.year), normalization, and conversion to and from Period, DatetimeIndex.to_period and PeriodIndex.to_timestamp (GH11263)
  • Release the GIL on some rolling algos: rolling_median, rolling_mean, rolling_max, rolling_min, rolling_var, rolling_kurt, rolling_skew (GH11450)
  • Release the GIL when reading and parsing text files in read_csv, read_table (GH11272)
  • Improved performance of rolling_median (GH11450)
  • Improved performance of to_excel (GH11352)
  • Performance bug in repr of Categorical categories, which was rendering the strings before chopping them for display (GH11305)
  • Performance improvement in Categorical.remove_unused_categories, (GH11643).
  • Improved performance of Series constructor with no data and DatetimeIndex (GH11433)
  • Improved performance of shift, cumprod, and cumsum with groupby (GH4095)

Bug Fixes

  • SparseArray.__iter__() now does not cause PendingDeprecationWarning in Python 3.5 (GH11622)
  • Regression from 0.16.2 for output formatting of long floats/nan, restored in (GH11302)
  • Series.sort_index() now correctly handles the inplace option (GH11402)
  • Incorrectly distributed .c file in the build on PyPi when reading a csv of floats and passing na_values=<a scalar> would show an exception (GH11374)
  • Bug in .to_latex() output broken when the index has a name (GH10660)
  • Bug in HDFStore.append with strings whose encoded length exceded the max unencoded length (GH11234)
  • Bug in merging datetime64[ns, tz] dtypes (GH11405)
  • Bug in HDFStore.select when comparing with a numpy scalar in a where clause (GH11283)
  • Bug in using DataFrame.ix with a multi-index indexer (GH11372)
  • Bug in date_range with ambigous endpoints (GH11626)
  • Prevent adding new attributes to the accessors .str, .dt and .cat. Retrieving such a value was not possible, so error out on setting it. (GH10673)
  • Bug in tz-conversions with an ambiguous time and .dt accessors (GH11295)
  • Bug in output formatting when using an index of ambiguous times (GH11619)
  • Bug in comparisons of Series vs list-likes (GH11339)
  • Bug in DataFrame.replace with a datetime64[ns, tz] and a non-compat to_replace (GH11326, GH11153)
  • Bug in isnull where numpy.datetime64('NaT') in a numpy.array was not determined to be null(GH11206)
  • Bug in list-like indexing with a mixed-integer Index (GH11320)
  • Bug in pivot_table with margins=True when indexes are of Categorical dtype (GH10993)
  • Bug in DataFrame.plot cannot use hex strings colors (GH10299)
  • Regression in DataFrame.drop_duplicates from 0.16.2, causing incorrect results on integer values (GH11376)
  • Bug in pd.eval where unary ops in a list error (GH11235)
  • Bug in squeeze() with zero length arrays (GH11230, GH8999)
  • Bug in describe() dropping column names for hierarchical indexes (GH11517)
  • Bug in DataFrame.pct_change() not propagating axis keyword on .fillna method (GH11150)
  • Bug in .to_csv() when a mix of integer and string column names are passed as the columns parameter (GH11637)
  • Bug in indexing with a range, (GH11652)
  • Bug in inference of numpy scalars and preserving dtype when setting columns (GH11638)
  • Bug in to_sql using unicode column names giving UnicodeEncodeError with (GH11431).
  • Fix regression in setting of xticks in plot (GH11529).
  • Bug in holiday.dates where observance rules could not be applied to holiday and doc enhancement (GH11477, GH11533)
  • Fix plotting issues when having plain Axes instances instead of SubplotAxes (GH11520, GH11556).
  • Bug in DataFrame.to_latex() produces an extra rule when header=False (GH7124)
  • Bug in df.groupby(...).apply(func) when a func returns a Series containing a new datetimelike column (GH11324)
  • Bug in pandas.json when file to load is big (GH11344)
  • Bugs in to_excel with duplicate columns (GH11007, GH10982, GH10970)
  • Fixed a bug that prevented the construction of an empty series of dtype datetime64[ns, tz] (GH11245).
  • Bug in read_excel with multi-index containing integers (GH11317)
  • Bug in to_excel with openpyxl 2.2+ and merging (GH11408)
  • Bug in DataFrame.to_dict() produces a np.datetime64 object instead of Timestamp when only datetime is present in data (GH11327)
  • Bug in DataFrame.corr() raises exception when computes Kendall correlation for DataFrames with boolean and not boolean columns (GH11560)
  • Bug in the link-time error caused by C inline functions on FreeBSD 10+ (with clang) (GH10510)
  • Bug in DataFrame.to_csv in passing through arguments for formatting MultiIndexes, including date_format (GH7791)
  • Bug in DataFrame.join() with how='right' producing a TypeError (GH11519)
  • Bug in Series.quantile with empty list results has Index with object dtype (GH11588)
  • Bug in pd.merge results in empty Int64Index rather than Index(dtype=object) when the merge result is empty (GH11588)
  • Bug in Categorical.remove_unused_categories when having NaN values (GH11599)
  • Bug in DataFrame.to_sparse() loses column names for MultiIndexes (GH11600)
  • Bug in DataFrame.round() with non-unique column index producing a Fatal Python error (GH11611)
  • Bug in DataFrame.round() with decimals being a non-unique indexed Series producing extra columns (GH11618)

v0.17.0 (October 9, 2015)

This is a major release from 0.16.2 and includes a small number of API changes, several new features, enhancements, and performance improvements along with a large number of bug fixes. We recommend that all users upgrade to this version.

Warning

pandas >= 0.17.0 will no longer support compatibility with Python version 3.2 (GH9118)

Warning

The pandas.io.data package is deprecated and will be replaced by the pandas-datareader package. This will allow the data modules to be independently updated to your pandas installation. The API for pandas-datareader v0.1.1 is exactly the same as in pandas v0.17.0 (GH8961, GH10861).

After installing pandas-datareader, you can easily change your imports:

from pandas.io import data, wb

becomes

from pandas_datareader import data, wb

Highlights include:

  • Release the Global Interpreter Lock (GIL) on some cython operations, see here
  • Plotting methods are now available as attributes of the .plot accessor, see here
  • The sorting API has been revamped to remove some long-time inconsistencies, see here
  • Support for a datetime64[ns] with timezones as a first-class dtype, see here
  • The default for to_datetime will now be to raise when presented with unparseable formats, previously this would return the original input. Also, date parse functions now return consistent results. See here
  • The default for dropna in HDFStore has changed to False, to store by default all rows even if they are all NaN, see here
  • Datetime accessor (dt) now supports Series.dt.strftime to generate formatted strings for datetime-likes, and Series.dt.total_seconds to generate each duration of the timedelta in seconds. See here
  • Period and PeriodIndex can handle multiplied freq like 3D, which corresponding to 3 days span. See here
  • Development installed versions of pandas will now have PEP440 compliant version strings (GH9518)
  • Development support for benchmarking with the Air Speed Velocity library (GH8361)
  • Support for reading SAS xport files, see here
  • Documentation comparing SAS to pandas, see here
  • Removal of the automatic TimeSeries broadcasting, deprecated since 0.8.0, see here
  • Display format with plain text can optionally align with Unicode East Asian Width, see here
  • Compatibility with Python 3.5 (GH11097)
  • Compatibility with matplotlib 1.5.0 (GH11111)

Check the API Changes and deprecations before updating.

New features

Datetime with TZ

We are adding an implementation that natively supports datetime with timezones. A Series or a DataFrame column previously could be assigned a datetime with timezones, and would work as an object dtype. This had performance issues with a large number rows. See the docs for more details. (GH8260, GH10763, GH11034).

The new implementation allows for having a single-timezone across all rows, with operations in a performant manner.

In [1]: df = DataFrame({'A' : date_range('20130101',periods=3),
   ...:                 'B' : date_range('20130101',periods=3,tz='US/Eastern'),
   ...:                 'C' : date_range('20130101',periods=3,tz='CET')})
   ...: 

In [2]: df
Out[2]: 
           A                         B                         C
0 2013-01-01 2013-01-01 00:00:00-05:00 2013-01-01 00:00:00+01:00
1 2013-01-02 2013-01-02 00:00:00-05:00 2013-01-02 00:00:00+01:00
2 2013-01-03 2013-01-03 00:00:00-05:00 2013-01-03 00:00:00+01:00

In [3]: df.dtypes
Out[3]: 
A                datetime64[ns]
B    datetime64[ns, US/Eastern]
C           datetime64[ns, CET]
dtype: object
In [4]: df.B
Out[4]: 
0   2013-01-01 00:00:00-05:00
1   2013-01-02 00:00:00-05:00
2   2013-01-03 00:00:00-05:00
Name: B, dtype: datetime64[ns, US/Eastern]

In [5]: df.B.dt.tz_localize(None)
Out[5]: 
0   2013-01-01
1   2013-01-02
2   2013-01-03
Name: B, dtype: datetime64[ns]

This uses a new-dtype representation as well, that is very similar in look-and-feel to its numpy cousin datetime64[ns]

In [6]: df['B'].dtype
Out[6]: datetime64[ns, US/Eastern]

In [7]: type(df['B'].dtype)
Out[7]: pandas.types.dtypes.DatetimeTZDtype

Note

There is a slightly different string repr for the underlying DatetimeIndex as a result of the dtype changes, but functionally these are the same.

Previous Behavior:

In [1]: pd.date_range('20130101',periods=3,tz='US/Eastern')
Out[1]: DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 00:00:00-05:00',
                       '2013-01-03 00:00:00-05:00'],
                      dtype='datetime64[ns]', freq='D', tz='US/Eastern')

In [2]: pd.date_range('20130101',periods=3,tz='US/Eastern').dtype
Out[2]: dtype('<M8[ns]')

New Behavior:

In [8]: pd.date_range('20130101',periods=3,tz='US/Eastern')
Out[8]: 
DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 00:00:00-05:00',
               '2013-01-03 00:00:00-05:00'],
              dtype='datetime64[ns, US/Eastern]', freq='D')

In [9]: pd.date_range('20130101',periods=3,tz='US/Eastern').dtype
Out[9]: datetime64[ns, US/Eastern]

Releasing the GIL

We are releasing the global-interpreter-lock (GIL) on some cython operations. This will allow other threads to run simultaneously during computation, potentially allowing performance improvements from multi-threading. Notably groupby, nsmallest, value_counts and some indexing operations benefit from this. (GH8882)

For example the groupby expression in the following code will have the GIL released during the factorization step, e.g. df.groupby('key') as well as the .sum() operation.

N = 1000000
ngroups = 10
df = DataFrame({'key' : np.random.randint(0,ngroups,size=N),
                'data' : np.random.randn(N) })
df.groupby('key')['data'].sum()

Releasing of the GIL could benefit an application that uses threads for user interactions (e.g. QT), or performing multi-threaded computations. A nice example of a library that can handle these types of computation-in-parallel is the dask library.

Plot submethods

The Series and DataFrame .plot() method allows for customizing plot types by supplying the kind keyword arguments. Unfortunately, many of these kinds of plots use different required and optional keyword arguments, which makes it difficult to discover what any given plot kind uses out of the dozens of possible arguments.

To alleviate this issue, we have added a new, optional plotting interface, which exposes each kind of plot as a method of the .plot attribute. Instead of writing series.plot(kind=<kind>, ...), you can now also use series.plot.<kind>(...):

In [10]: df = pd.DataFrame(np.random.rand(10, 2), columns=['a', 'b'])

In [11]: df.plot.bar()
_images/whatsnew_plot_submethods.png

As a result of this change, these methods are now all discoverable via tab-completion:

In [12]: df.plot.<TAB>
df.plot.area     df.plot.barh     df.plot.density  df.plot.hist     df.plot.line     df.plot.scatter
df.plot.bar      df.plot.box      df.plot.hexbin   df.plot.kde      df.plot.pie

Each method signature only includes relevant arguments. Currently, these are limited to required arguments, but in the future these will include optional arguments, as well. For an overview, see the new Plotting API documentation.

Additional methods for dt accessor

strftime

We are now supporting a Series.dt.strftime method for datetime-likes to generate a formatted string (GH10110). Examples:

# DatetimeIndex
In [13]: s = pd.Series(pd.date_range('20130101', periods=4))

In [14]: s
Out[14]: 
0   2013-01-01
1   2013-01-02
2   2013-01-03
3   2013-01-04
dtype: datetime64[ns]

In [15]: s.dt.strftime('%Y/%m/%d')
Out[15]: 
0    2013/01/01
1    2013/01/02
2    2013/01/03
3    2013/01/04
dtype: object
# PeriodIndex
In [16]: s = pd.Series(pd.period_range('20130101', periods=4))

In [17]: s
Out[17]: 
0   2013-01-01
1   2013-01-02
2   2013-01-03
3   2013-01-04
dtype: object

In [18]: s.dt.strftime('%Y/%m/%d')
Out[18]: 
0    2013/01/01
1    2013/01/02
2    2013/01/03
3    2013/01/04
dtype: object

The string format is as the python standard library and details can be found here

total_seconds

pd.Series of type timedelta64 has new method .dt.total_seconds() returning the duration of the timedelta in seconds (GH10817)

# TimedeltaIndex
In [19]: s = pd.Series(pd.timedelta_range('1 minutes', periods=4))

In [20]: s
Out[20]: 
0   0 days 00:01:00
1   1 days 00:01:00
2   2 days 00:01:00
3   3 days 00:01:00
dtype: timedelta64[ns]

In [21]: s.dt.total_seconds()
Out[21]: 
0        60.0
1     86460.0
2    172860.0
3    259260.0
dtype: float64

Period Frequency Enhancement

Period, PeriodIndex and period_range can now accept multiplied freq. Also, Period.freq and PeriodIndex.freq are now stored as a DateOffset instance like DatetimeIndex, and not as str (GH7811)

A multiplied freq represents a span of corresponding length. The example below creates a period of 3 days. Addition and subtraction will shift the period by its span.

In [22]: p = pd.Period('2015-08-01', freq='3D')

In [23]: p
Out[23]: Period('2015-08-01', '3D')

In [24]: p + 1
Out[24]: Period('2015-08-04', '3D')

In [25]: p - 2
Out[25]: Period('2015-07-26', '3D')

In [26]: p.to_timestamp()
Out[26]: Timestamp('2015-08-01 00:00:00')

In [27]: p.to_timestamp(how='E')
Out[27]: Timestamp('2015-08-03 00:00:00')

You can use the multiplied freq in PeriodIndex and period_range.

In [28]: idx = pd.period_range('2015-08-01', periods=4, freq='2D')

In [29]: idx
Out[29]: PeriodIndex(['2015-08-01', '2015-08-03', '2015-08-05', '2015-08-07'], dtype='period[2D]', freq='2D')

In [30]: idx + 1
Out[30]: PeriodIndex(['2015-08-03', '2015-08-05', '2015-08-07', '2015-08-09'], dtype='period[2D]', freq='2D')

Support for SAS XPORT files

read_sas() provides support for reading SAS XPORT format files. (GH4052).

df = pd.read_sas('sas_xport.xpt')

It is also possible to obtain an iterator and read an XPORT file incrementally.

for df in pd.read_sas('sas_xport.xpt', chunksize=10000)
    do_something(df)

See the docs for more details.

Support for Math Functions in .eval()

eval() now supports calling math functions (GH4893)

df = pd.DataFrame({'a': np.random.randn(10)})
df.eval("b = sin(a)")

The support math functions are sin, cos, exp, log, expm1, log1p, sqrt, sinh, cosh, tanh, arcsin, arccos, arctan, arccosh, arcsinh, arctanh, abs and arctan2.

These functions map to the intrinsics for the NumExpr engine. For the Python engine, they are mapped to NumPy calls.

Changes to Excel with MultiIndex

In version 0.16.2 a DataFrame with MultiIndex columns could not be written to Excel via to_excel. That functionality has been added (GH10564), along with updating read_excel so that the data can be read back with, no loss of information, by specifying which columns/rows make up the MultiIndex in the header and index_col parameters (GH4679)

See the documentation for more details.

In [31]: df = pd.DataFrame([[1,2,3,4], [5,6,7,8]],
   ....:                   columns = pd.MultiIndex.from_product([['foo','bar'],['a','b']],
   ....:                                                        names = ['col1', 'col2']),
   ....:                   index = pd.MultiIndex.from_product([['j'], ['l', 'k']],
   ....:                                                      names = ['i1', 'i2']))
   ....: 

In [32]: df
Out[32]: 
col1  foo    bar   
col2    a  b   a  b
i1 i2              
j  l    1  2   3  4
   k    5  6   7  8

In [33]: df.to_excel('test.xlsx')

In [34]: df = pd.read_excel('test.xlsx', header=[0,1], index_col=[0,1])

In [35]: df
Out[35]: 
col1  foo    bar   
col2    a  b   a  b
i1 i2              
j  l    1  2   3  4
   k    5  6   7  8

Previously, it was necessary to specify the has_index_names argument in read_excel, if the serialized data had index names. For version 0.17.0 the ouptput format of to_excel has been changed to make this keyword unnecessary - the change is shown below.

Old

_images/old-excel-index.png

New

_images/new-excel-index.png

Warning

Excel files saved in version 0.16.2 or prior that had index names will still able to be read in, but the has_index_names argument must specified to True.

Google BigQuery Enhancements

  • Added ability to automatically create a table/dataset using the pandas.io.gbq.to_gbq() function if the destination table/dataset does not exist. (GH8325, GH11121).
  • Added ability to replace an existing table and schema when calling the pandas.io.gbq.to_gbq() function via the if_exists argument. See the docs for more details (GH8325).
  • InvalidColumnOrder and InvalidPageToken in the gbq module will raise ValueError instead of IOError.
  • The generate_bq_schema() function is now deprecated and will be removed in a future version (GH11121)
  • The gbq module will now support Python 3 (GH11094).

Display Alignment with Unicode East Asian Width

Warning

Enabling this option will affect the performance for printing of DataFrame and Series (about 2 times slower). Use only when it is actually required.

Some East Asian countries use Unicode characters its width is corresponding to 2 alphabets. If a DataFrame or Series contains these characters, the default output cannot be aligned properly. The following options are added to enable precise handling for these characters.

  • display.unicode.east_asian_width: Whether to use the Unicode East Asian Width to calculate the display text width. (GH2612)
  • display.unicode.ambiguous_as_wide: Whether to handle Unicode characters belong to Ambiguous as Wide. (GH11102)
In [36]: df = pd.DataFrame({u'国籍': ['UK', u'日本'], u'名前': ['Alice', u'しのぶ']})

In [37]: df;
_images/option_unicode01.png
In [38]: pd.set_option('display.unicode.east_asian_width', True)

In [39]: df;
_images/option_unicode02.png

For further details, see here

Other enhancements

  • Support for openpyxl >= 2.2. The API for style support is now stable (GH10125)

  • merge now accepts the argument indicator which adds a Categorical-type column (by default called _merge) to the output object that takes on the values (GH8790)

    Observation Origin _merge value
    Merge key only in 'left' frame left_only
    Merge key only in 'right' frame right_only
    Merge key in both frames both
    In [40]: df1 = pd.DataFrame({'col1':[0,1], 'col_left':['a','b']})
    
    In [41]: df2 = pd.DataFrame({'col1':[1,2,2],'col_right':[2,2,2]})
    
    In [42]: pd.merge(df1, df2, on='col1', how='outer', indicator=True)
    Out[42]: 
       col1 col_left  col_right      _merge
    0     0        a        NaN   left_only
    1     1        b        2.0        both
    2     2      NaN        2.0  right_only
    3     2      NaN        2.0  right_only
    

    For more, see the updated docs

  • pd.to_numeric is a new function to coerce strings to numbers (possibly with coercion) (GH11133)

  • pd.merge will now allow duplicate column names if they are not merged upon (GH10639).

  • pd.pivot will now allow passing index as None (GH3962).

  • pd.concat will now use existing Series names if provided (GH10698).

    In [43]: foo = pd.Series([1,2], name='foo')
    
    In [44]: bar = pd.Series([1,2])
    
    In [45]: baz = pd.Series([4,5])
    

    Previous Behavior:

    In [1] pd.concat([foo, bar, baz], 1)
    Out[1]:
          0  1  2
       0  1  1  4
       1  2  2  5
    

    New Behavior:

    In [46]: pd.concat([foo, bar, baz], 1)
    Out[46]: 
       foo  0  1
    0    1  1  4
    1    2  2  5
    
  • DataFrame has gained the nlargest and nsmallest methods (GH10393)

  • Add a limit_direction keyword argument that works with limit to enable interpolate to fill NaN values forward, backward, or both (GH9218, GH10420, GH11115)

    In [47]: ser = pd.Series([np.nan, np.nan, 5, np.nan, np.nan, np.nan, 13])
    
    In [48]: ser.interpolate(limit=1, limit_direction='both')
    Out[48]: 
    0     NaN
    1     5.0
    2     5.0
    3     7.0
    4     NaN
    5    11.0
    6    13.0
    dtype: float64
    
  • Added a DataFrame.round method to round the values to a variable number of decimal places (GH10568).

    In [49]: df = pd.DataFrame(np.random.random([3, 3]), columns=['A', 'B', 'C'],
       ....: index=['first', 'second', 'third'])
       ....: 
    
    In [50]: df
    Out[50]: 
                   A         B         C
    first   0.342764  0.304121  0.417022
    second  0.681301  0.875457  0.510422
    third   0.669314  0.585937  0.624904
    
    In [51]: df.round(2)
    Out[51]: 
               A     B     C
    first   0.34  0.30  0.42
    second  0.68  0.88  0.51
    third   0.67  0.59  0.62
    
    In [52]: df.round({'A': 0, 'C': 2})
    Out[52]: 
              A         B     C
    first   0.0  0.304121  0.42
    second  1.0  0.875457  0.51
    third   1.0  0.585937  0.62
    
  • drop_duplicates and duplicated now accept a keep keyword to target first, last, and all duplicates. The take_last keyword is deprecated, see here (GH6511, GH8505)

    In [53]: s = pd.Series(['A', 'B', 'C', 'A', 'B', 'D'])
    
    In [54]: s.drop_duplicates()
    Out[54]: 
    0    A
    1    B
    2    C
    5    D
    dtype: object
    
    In [55]: s.drop_duplicates(keep='last')
    Out[55]: 
    2    C
    3    A
    4    B
    5    D
    dtype: object
    
    In [56]: s.drop_duplicates(keep=False)
    Out[56]: 
    2    C
    5    D
    dtype: object
    
  • Reindex now has a tolerance argument that allows for finer control of Limits on filling while reindexing (GH10411):

    In [57]: df = pd.DataFrame({'x': range(5),
       ....:                    't': pd.date_range('2000-01-01', periods=5)})
       ....: 
    
    In [58]: df.reindex([0.1, 1.9, 3.5],
       ....:            method='nearest',
       ....:            tolerance=0.2)
       ....: 
    Out[58]: 
                 t    x
    0.1 2000-01-01  0.0
    1.9 2000-01-03  2.0
    3.5        NaT  NaN
    

    When used on a DatetimeIndex, TimedeltaIndex or PeriodIndex, tolerance will coerced into a Timedelta if possible. This allows you to specify tolerance with a string:

    In [59]: df = df.set_index('t')
    
    In [60]: df.reindex(pd.to_datetime(['1999-12-31']),
       ....:            method='nearest',
       ....:            tolerance='1 day')
       ....: 
    Out[60]: 
                x
    1999-12-31  0
    

    tolerance is also exposed by the lower level Index.get_indexer and Index.get_loc methods.

  • Added functionality to use the base argument when resampling a TimeDeltaIndex (GH10530)

  • DatetimeIndex can be instantiated using strings contains NaT (GH7599)

  • to_datetime can now accept the yearfirst keyword (GH7599)

  • pandas.tseries.offsets larger than the Day offset can now be used with a Series for addition/subtraction (GH10699). See the docs for more details.

  • pd.Timedelta.total_seconds() now returns Timedelta duration to ns precision (previously microsecond precision) (GH10939)

  • PeriodIndex now supports arithmetic with np.ndarray (GH10638)

  • Support pickling of Period objects (GH10439)

  • .as_blocks will now take a copy optional argument to return a copy of the data, default is to copy (no change in behavior from prior versions), (GH9607)

  • regex argument to DataFrame.filter now handles numeric column names instead of raising ValueError (GH10384).

  • Enable reading gzip compressed files via URL, either by explicitly setting the compression parameter or by inferring from the presence of the HTTP Content-Encoding header in the response (GH8685)

  • Enable writing Excel files in memory using StringIO/BytesIO (GH7074)

  • Enable serialization of lists and dicts to strings in ExcelWriter (GH8188)

  • SQL io functions now accept a SQLAlchemy connectable. (GH7877)

  • pd.read_sql and to_sql can accept database URI as con parameter (GH10214)

  • read_sql_table will now allow reading from views (GH10750).

  • Enable writing complex values to HDFStores when using the table format (GH10447)

  • Enable pd.read_hdf to be used without specifying a key when the HDF file contains a single dataset (GH10443)

  • pd.read_stata will now read Stata 118 type files. (GH9882)

  • msgpack submodule has been updated to 0.4.6 with backward compatibility (GH10581)

  • DataFrame.to_dict now accepts orient='index' keyword argument (GH10844).

  • DataFrame.apply will return a Series of dicts if the passed function returns a dict and reduce=True (GH8735).

  • Allow passing kwargs to the interpolation methods (GH10378).

  • Improved error message when concatenating an empty iterable of Dataframe objects (GH9157)

  • pd.read_csv can now read bz2-compressed files incrementally, and the C parser can read bz2-compressed files from AWS S3 (GH11070, GH11072).

  • In pd.read_csv, recognize s3n:// and s3a:// URLs as designating S3 file storage (GH11070, GH11071).

  • Read CSV files from AWS S3 incrementally, instead of first downloading the entire file. (Full file download still required for compressed files in Python 2.) (GH11070, GH11073)

  • pd.read_csv is now able to infer compression type for files read from AWS S3 storage (GH11070, GH11074).

Backwards incompatible API changes

Changes to sorting API

The sorting API has had some longtime inconsistencies. (GH9816, GH8239).

Here is a summary of the API PRIOR to 0.17.0:

  • Series.sort is INPLACE while DataFrame.sort returns a new object.
  • Series.order returns a new object
  • It was possible to use Series/DataFrame.sort_index to sort by values by passing the by keyword.
  • Series/DataFrame.sortlevel worked only on a MultiIndex for sorting by index.

To address these issues, we have revamped the API:

  • We have introduced a new method, DataFrame.sort_values(), which is the merger of DataFrame.sort(), Series.sort(), and Series.order(), to handle sorting of values.
  • The existing methods Series.sort(), Series.order(), and DataFrame.sort() have been deprecated and will be removed in a future version.
  • The by argument of DataFrame.sort_index() has been deprecated and will be removed in a future version.
  • The existing method .sort_index() will gain the level keyword to enable level sorting.

We now have two distinct and non-overlapping methods of sorting. A * marks items that will show a FutureWarning.

To sort by the values:

Previous Replacement
* Series.order() Series.sort_values()
* Series.sort() Series.sort_values(inplace=True)
* DataFrame.sort(columns=...) DataFrame.sort_values(by=...)

To sort by the index:

Previous Replacement
Series.sort_index() Series.sort_index()
Series.sortlevel(level=...) Series.sort_index(level=...)
DataFrame.sort_index() DataFrame.sort_index()
DataFrame.sortlevel(level=...) DataFrame.sort_index(level=...)
* DataFrame.sort() DataFrame.sort_index()

We have also deprecated and changed similar methods in two Series-like classes, Index and Categorical.

Previous Replacement
* Index.order() Index.sort_values()
* Categorical.order() Categorical.sort_values()

Changes to to_datetime and to_timedelta

Error handling

The default for pd.to_datetime error handling has changed to errors='raise'. In prior versions it was errors='ignore'. Furthermore, the coerce argument has been deprecated in favor of errors='coerce'. This means that invalid parsing will raise rather that return the original input as in previous versions. (GH10636)

Previous Behavior:

In [2]: pd.to_datetime(['2009-07-31', 'asd'])
Out[2]: array(['2009-07-31', 'asd'], dtype=object)

New Behavior:

In [3]: pd.to_datetime(['2009-07-31', 'asd'])
ValueError: Unknown string format

Of course you can coerce this as well.

In [61]: to_datetime(['2009-07-31', 'asd'], errors='coerce')
Out[61]: DatetimeIndex(['2009-07-31', 'NaT'], dtype='datetime64[ns]', freq=None)

To keep the previous behavior, you can use errors='ignore':

In [62]: to_datetime(['2009-07-31', 'asd'], errors='ignore')
Out[62]: array(['2009-07-31', 'asd'], dtype=object)

Furthermore, pd.to_timedelta has gained a similar API, of errors='raise'|'ignore'|'coerce', and the coerce keyword has been deprecated in favor of errors='coerce'.

Consistent Parsing

The string parsing of to_datetime, Timestamp and DatetimeIndex has been made consistent. (GH7599)

Prior to v0.17.0, Timestamp and to_datetime may parse year-only datetime-string incorrectly using today’s date, otherwise DatetimeIndex uses the beginning of the year. Timestamp and to_datetime may raise ValueError in some types of datetime-string which DatetimeIndex can parse, such as a quarterly string.

Previous Behavior:

In [1]: Timestamp('2012Q2')
Traceback
   ...
ValueError: Unable to parse 2012Q2

# Results in today's date.
In [2]: Timestamp('2014')
Out [2]: 2014-08-12 00:00:00

v0.17.0 can parse them as below. It works on DatetimeIndex also.

New Behavior:

In [63]: Timestamp('2012Q2')
Out[63]: Timestamp('2012-04-01 00:00:00')

In [64]: Timestamp('2014')
Out[64]: Timestamp('2014-01-01 00:00:00')

In [65]: DatetimeIndex(['2012Q2', '2014'])
Out[65]: DatetimeIndex(['2012-04-01', '2014-01-01'], dtype='datetime64[ns]', freq=None)

Note

If you want to perform calculations based on today’s date, use Timestamp.now() and pandas.tseries.offsets.

In [66]: import pandas.tseries.offsets as offsets

In [67]: Timestamp.now()
Out[67]: Timestamp('2016-11-03 16:51:06.549337')

In [68]: Timestamp.now() + offsets.DateOffset(years=1)
Out[68]: Timestamp('2017-11-03 16:51:06.550998')

Changes to Index Comparisons

Operator equal on Index should behavior similarly to Series (GH9947, GH10637)

Starting in v0.17.0, comparing Index objects of different lengths will raise a ValueError. This is to be consistent with the behavior of Series.

Previous Behavior:

In [2]: pd.Index([1, 2, 3]) == pd.Index([1, 4, 5])
Out[2]: array([ True, False, False], dtype=bool)

In [3]: pd.Index([1, 2, 3]) == pd.Index([2])
Out[3]: array([False,  True, False], dtype=bool)

In [4]: pd.Index([1, 2, 3]) == pd.Index([1, 2])
Out[4]: False

New Behavior:

In [8]: pd.Index([1, 2, 3]) == pd.Index([1, 4, 5])
Out[8]: array([ True, False, False], dtype=bool)

In [9]: pd.Index([1, 2, 3]) == pd.Index([2])
ValueError: Lengths must match to compare

In [10]: pd.Index([1, 2, 3]) == pd.Index([1, 2])
ValueError: Lengths must match to compare

Note that this is different from the numpy behavior where a comparison can be broadcast:

In [69]: np.array([1, 2, 3]) == np.array([1])
Out[69]: array([ True, False, False], dtype=bool)

or it can return False if broadcasting can not be done:

In [70]: np.array([1, 2, 3]) == np.array([1, 2])
Out[70]: False

Changes to Boolean Comparisons vs. None

Boolean comparisons of a Series vs None will now be equivalent to comparing with np.nan, rather than raise TypeError. (GH1079).

In [71]: s = Series(range(3))

In [72]: s.iloc[1] = None

In [73]: s
Out[73]: 
0    0.0
1    NaN
2    2.0
dtype: float64

Previous Behavior:

In [5]: s==None
TypeError: Could not compare <type 'NoneType'> type with Series

New Behavior:

In [74]: s==None
Out[74]: 
0    False
1    False
2    False
dtype: bool

Usually you simply want to know which values are null.

In [75]: s.isnull()
Out[75]: 
0    False
1     True
2    False
dtype: bool

Warning

You generally will want to use isnull/notnull for these types of comparisons, as isnull/notnull tells you which elements are null. One has to be mindful that nan's don’t compare equal, but None's do. Note that Pandas/numpy uses the fact that np.nan != np.nan, and treats None like np.nan.

In [76]: None == None
Out[76]: True

In [77]: np.nan == np.nan
Out[77]: False

HDFStore dropna behavior

The default behavior for HDFStore write functions with format='table' is now to keep rows that are all missing. Previously, the behavior was to drop rows that were all missing save the index. The previous behavior can be replicated using the dropna=True option. (GH9382)

Previous Behavior:

In [78]: df_with_missing = pd.DataFrame({'col1':[0, np.nan, 2],
   ....:                                 'col2':[1, np.nan, np.nan]})
   ....: 

In [79]: df_with_missing
Out[79]: 
   col1  col2
0   0.0   1.0
1   NaN   NaN
2   2.0   NaN
In [27]:
df_with_missing.to_hdf('file.h5',
                       'df_with_missing',
                       format='table',
                       mode='w')

In [28]: pd.read_hdf('file.h5', 'df_with_missing')

Out [28]:
      col1  col2
  0     0     1
  2     2   NaN

New Behavior:

In [80]: df_with_missing.to_hdf('file.h5',
   ....:                        'df_with_missing',
   ....:                         format='table',
   ....:                         mode='w')
   ....: 

In [81]: pd.read_hdf('file.h5', 'df_with_missing')
Out[81]: 
   col1  col2
0   0.0   1.0
1   NaN   NaN
2   2.0   NaN

See the docs for more details.

Changes to display.precision option

The display.precision option has been clarified to refer to decimal places (GH10451).

Earlier versions of pandas would format floating point numbers to have one less decimal place than the value in display.precision.

In [1]: pd.set_option('display.precision', 2)

In [2]: pd.DataFrame({'x': [123.456789]})
Out[2]:
       x
0  123.5

If interpreting precision as “significant figures” this did work for scientific notation but that same interpretation did not work for values with standard formatting. It was also out of step with how numpy handles formatting.

Going forward the value of display.precision will directly control the number of places after the decimal, for regular formatting as well as scientific notation, similar to how numpy’s precision print option works.

In [82]: pd.set_option('display.precision', 2)

In [83]: pd.DataFrame({'x': [123.456789]})
Out[83]: 
        x
0  123.46

To preserve output behavior with prior versions the default value of display.precision has been reduced to 6 from 7.

Changes to Categorical.unique

Categorical.unique now returns new Categoricals with categories and codes that are unique, rather than returning np.array (GH10508)

  • unordered category: values and categories are sorted by appearance order.
  • ordered category: values are sorted by appearance order, categories keep existing order.
In [84]: cat = pd.Categorical(['C', 'A', 'B', 'C'],
   ....:                      categories=['A', 'B', 'C'],
   ....:                      ordered=True)
   ....: 

In [85]: cat
Out[85]: 
[C, A, B, C]
Categories (3, object): [A < B < C]

In [86]: cat.unique()
Out[86]: 
[C, A, B]
Categories (3, object): [A < B < C]

In [87]: cat = pd.Categorical(['C', 'A', 'B', 'C'],
   ....:                      categories=['A', 'B', 'C'])
   ....: 

In [88]: cat
Out[88]: 
[C, A, B, C]
Categories (3, object): [A, B, C]

In [89]: cat.unique()
Out[89]: 
[C, A, B]
Categories (3, object): [C, A, B]

Changes to bool passed as header in Parsers

In earlier versions of pandas, if a bool was passed the header argument of read_csv, read_excel, or read_html it was implicitly converted to an integer, resulting in header=0 for False and header=1 for True (GH6113)

A bool input to header will now raise a TypeError

In [29]: df = pd.read_csv('data.csv', header=False)
TypeError: Passing a bool to header is invalid. Use header=None for no header or
header=int or list-like of ints to specify the row(s) making up the column names

Other API Changes

  • Line and kde plot with subplots=True now uses default colors, not all black. Specify color='k' to draw all lines in black (GH9894)

  • Calling the .value_counts() method on a Series with a categorical dtype now returns a Series with a CategoricalIndex (GH10704)

  • The metadata properties of subclasses of pandas objects will now be serialized (GH10553).

  • groupby using Categorical follows the same rule as Categorical.unique described above (GH10508)

  • When constructing DataFrame with an array of complex64 dtype previously meant the corresponding column was automatically promoted to the complex128 dtype. Pandas will now preserve the itemsize of the input for complex data (GH10952)

  • some numeric reduction operators would return ValueError, rather than TypeError on object types that includes strings and numbers (GH11131)

  • Passing currently unsupported chunksize argument to read_excel or ExcelFile.parse will now raise NotImplementedError (GH8011)

  • Allow an ExcelFile object to be passed into read_excel (GH11198)

  • DatetimeIndex.union does not infer freq if self and the input have None as freq (GH11086)

  • NaT‘s methods now either raise ValueError, or return np.nan or NaT (GH9513)

    Behavior Methods
    return np.nan weekday, isoweekday
    return NaT date, now, replace, to_datetime, today
    return np.datetime64('NaT') to_datetime64 (unchanged)
    raise ValueError All other public methods (names not beginning with underscores)

Deprecations

  • For Series the following indexing functions are deprecated (GH10177).

    Deprecated Function Replacement
    .irow(i) .iloc[i] or .iat[i]
    .iget(i) .iloc[i] or .iat[i]
    .iget_value(i) .iloc[i] or .iat[i]
  • For DataFrame the following indexing functions are deprecated (GH10177).

    Deprecated Function Replacement
    .irow(i) .iloc[i]
    .iget_value(i, j) .iloc[i, j] or .iat[i, j]
    .icol(j) .iloc[:, j]

Note

These indexing function have been deprecated in the documentation since 0.11.0.

  • Categorical.name was deprecated to make Categorical more numpy.ndarray like. Use Series(cat, name="whatever") instead (GH10482).
  • Setting missing values (NaN) in a Categorical‘s categories will issue a warning (GH10748). You can still have missing values in the values.
  • drop_duplicates and duplicated‘s take_last keyword was deprecated in favor of keep. (GH6511, GH8505)
  • Series.nsmallest and nlargest‘s take_last keyword was deprecated in favor of keep. (GH10792)
  • DataFrame.combineAdd and DataFrame.combineMult are deprecated. They can easily be replaced by using the add and mul methods: DataFrame.add(other, fill_value=0) and DataFrame.mul(other, fill_value=1.) (GH10735).
  • TimeSeries deprecated in favor of Series (note that this has been an alias since 0.13.0), (GH10890)
  • SparsePanel deprecated and will be removed in a future version (GH11157).
  • Series.is_time_series deprecated in favor of Series.index.is_all_dates (GH11135)
  • Legacy offsets (like 'A@JAN') are deprecated (note that this has been alias since 0.8.0) (GH10878)
  • WidePanel deprecated in favor of Panel, LongPanel in favor of DataFrame (note these have been aliases since < 0.11.0), (GH10892)
  • DataFrame.convert_objects has been deprecated in favor of type-specific functions pd.to_datetime, pd.to_timestamp and pd.to_numeric (new in 0.17.0) (GH11133).

Removal of prior version deprecations/changes

  • Removal of na_last parameters from Series.order() and Series.sort(), in favor of na_position. (GH5231)

  • Remove of percentile_width from .describe(), in favor of percentiles. (GH7088)

  • Removal of colSpace parameter from DataFrame.to_string(), in favor of col_space, circa 0.8.0 version.

  • Removal of automatic time-series broadcasting (GH2304)

    In [90]: np.random.seed(1234)
    
    In [91]: df = DataFrame(np.random.randn(5,2),columns=list('AB'),index=date_range('20130101',periods=5))
    
    In [92]: df
    Out[92]: 
                       A         B
    2013-01-01  0.471435 -1.190976
    2013-01-02  1.432707 -0.312652
    2013-01-03 -0.720589  0.887163
    2013-01-04  0.859588 -0.636524
    2013-01-05  0.015696 -2.242685
    

    Previously

    In [3]: df + df.A
    FutureWarning: TimeSeries broadcasting along DataFrame index by default is deprecated.
    Please use DataFrame.<op> to explicitly broadcast arithmetic operations along the index
    
    Out[3]:
                        A         B
    2013-01-01  0.942870 -0.719541
    2013-01-02  2.865414  1.120055
    2013-01-03 -1.441177  0.166574
    2013-01-04  1.719177  0.223065
    2013-01-05  0.031393 -2.226989
    

    Current

    In [93]: df.add(df.A,axis='index')
    Out[93]: 
                       A         B
    2013-01-01  0.942870 -0.719541
    2013-01-02  2.865414  1.120055
    2013-01-03 -1.441177  0.166574
    2013-01-04  1.719177  0.223065
    2013-01-05  0.031393 -2.226989
    
  • Remove table keyword in HDFStore.put/append, in favor of using format= (GH4645)

  • Remove kind in read_excel/ExcelFile as its unused (GH4712)

  • Remove infer_type keyword from pd.read_html as its unused (GH4770, GH7032)

  • Remove offset and timeRule keywords from Series.tshift/shift, in favor of freq (GH4853, GH4864)

  • Remove pd.load/pd.save aliases in favor of pd.to_pickle/pd.read_pickle (GH3787)

Performance Improvements

  • Development support for benchmarking with the Air Speed Velocity library (GH8361)
  • Added vbench benchmarks for alternative ExcelWriter engines and reading Excel files (GH7171)
  • Performance improvements in Categorical.value_counts (GH10804)
  • Performance improvements in SeriesGroupBy.nunique and SeriesGroupBy.value_counts and SeriesGroupby.transform (GH10820, GH11077)
  • Performance improvements in DataFrame.drop_duplicates with integer dtypes (GH10917)
  • Performance improvements in DataFrame.duplicated with wide frames. (GH10161, GH11180)
  • 4x improvement in timedelta string parsing (GH6755, GH10426)
  • 8x improvement in timedelta64 and datetime64 ops (GH6755)
  • Significantly improved performance of indexing MultiIndex with slicers (GH10287)
  • 8x improvement in iloc using list-like input (GH10791)
  • Improved performance of Series.isin for datetimelike/integer Series (GH10287)
  • 20x improvement in concat of Categoricals when categories are identical (GH10587)
  • Improved performance of to_datetime when specified format string is ISO8601 (GH10178)
  • 2x improvement of Series.value_counts for float dtype (GH10821)
  • Enable infer_datetime_format in to_datetime when date components do not have 0 padding (GH11142)
  • Regression from 0.16.1 in constructing DataFrame from nested dictionary (GH11084)
  • Performance improvements in addition/subtraction operations for DateOffset with Series or DatetimeIndex (GH10744, GH11205)

Bug Fixes

  • Bug in incorrection computation of .mean() on timedelta64[ns] because of overflow (GH9442)
  • Bug in .isin on older numpies (:issue: 11232)
  • Bug in DataFrame.to_html(index=False) renders unnecessary name row (GH10344)
  • Bug in DataFrame.to_latex() the column_format argument could not be passed (GH9402)
  • Bug in DatetimeIndex when localizing with NaT (GH10477)
  • Bug in Series.dt ops in preserving meta-data (GH10477)
  • Bug in preserving NaT when passed in an otherwise invalid to_datetime construction (GH10477)
  • Bug in DataFrame.apply when function returns categorical series. (GH9573)
  • Bug in to_datetime with invalid dates and formats supplied (GH10154)
  • Bug in Index.drop_duplicates dropping name(s) (GH10115)
  • Bug in Series.quantile dropping name (GH10881)
  • Bug in pd.Series when setting a value on an empty Series whose index has a frequency. (GH10193)
  • Bug in pd.Series.interpolate with invalid order keyword values. (GH10633)
  • Bug in DataFrame.plot raises ValueError when color name is specified by multiple characters (GH10387)
  • Bug in Index construction with a mixed list of tuples (GH10697)
  • Bug in DataFrame.reset_index when index contains NaT. (GH10388)
  • Bug in ExcelReader when worksheet is empty (GH6403)
  • Bug in BinGrouper.group_info where returned values are not compatible with base class (GH10914)
  • Bug in clearing the cache on DataFrame.pop and a subsequent inplace op (GH10912)
  • Bug in indexing with a mixed-integer Index causing an ImportError (GH10610)
  • Bug in Series.count when index has nulls (GH10946)
  • Bug in pickling of a non-regular freq DatetimeIndex (GH11002)
  • Bug causing DataFrame.where to not respect the axis parameter when the frame has a symmetric shape. (GH9736)
  • Bug in Table.select_column where name is not preserved (GH10392)
  • Bug in offsets.generate_range where start and end have finer precision than offset (GH9907)
  • Bug in pd.rolling_* where Series.name would be lost in the output (GH10565)
  • Bug in stack when index or columns are not unique. (GH10417)
  • Bug in setting a Panel when an axis has a multi-index (GH10360)
  • Bug in USFederalHolidayCalendar where USMemorialDay and USMartinLutherKingJr were incorrect (GH10278 and GH9760 )
  • Bug in .sample() where returned object, if set, gives unnecessary SettingWithCopyWarning (GH10738)
  • Bug in .sample() where weights passed as Series were not aligned along axis before being treated positionally, potentially causing problems if weight indices were not aligned with sampled object. (GH10738)
  • Regression fixed in (GH9311, GH6620, GH9345), where groupby with a datetime-like converting to float with certain aggregators (GH10979)
  • Bug in DataFrame.interpolate with axis=1 and inplace=True (GH10395)
  • Bug in io.sql.get_schema when specifying multiple columns as primary key (GH10385).
  • Bug in groupby(sort=False) with datetime-like Categorical raises ValueError (GH10505)
  • Bug in groupby(axis=1) with filter() throws IndexError (GH11041)
  • Bug in test_categorical on big-endian builds (GH10425)
  • Bug in Series.shift and DataFrame.shift not supporting categorical data (GH9416)
  • Bug in Series.map using categorical Series raises AttributeError (GH10324)
  • Bug in MultiIndex.get_level_values including Categorical raises AttributeError (GH10460)
  • Bug in pd.get_dummies with sparse=True not returning SparseDataFrame (GH10531)
  • Bug in Index subtypes (such as PeriodIndex) not returning their own type for .drop and .insert methods (GH10620)
  • Bug in algos.outer_join_indexer when right array is empty (GH10618)
  • Bug in filter (regression from 0.16.0) and transform when grouping on multiple keys, one of which is datetime-like (GH10114)
  • Bug in to_datetime and to_timedelta causing Index name to be lost (GH10875)
  • Bug in len(DataFrame.groupby) causing IndexError when there’s a column containing only NaNs (:issue: 11016)
  • Bug that caused segfault when resampling an empty Series (GH10228)
  • Bug in DatetimeIndex and PeriodIndex.value_counts resets name from its result, but retains in result’s Index. (GH10150)
  • Bug in pd.eval using numexpr engine coerces 1 element numpy array to scalar (GH10546)
  • Bug in pd.concat with axis=0 when column is of dtype category (GH10177)
  • Bug in read_msgpack where input type is not always checked (GH10369, GH10630)
  • Bug in pd.read_csv with kwargs index_col=False, index_col=['a', 'b'] or dtype (GH10413, GH10467, GH10577)
  • Bug in Series.from_csv with header kwarg not setting the Series.name or the Series.index.name (GH10483)
  • Bug in groupby.var which caused variance to be inaccurate for small float values (GH10448)
  • Bug in Series.plot(kind='hist') Y Label not informative (GH10485)
  • Bug in read_csv when using a converter which generates a uint8 type (GH9266)
  • Bug causes memory leak in time-series line and area plot (GH9003)
  • Bug when setting a Panel sliced along the major or minor axes when the right-hand side is a DataFrame (GH11014)
  • Bug that returns None and does not raise NotImplementedError when operator functions (e.g. .add) of Panel are not implemented (GH7692)
  • Bug in line and kde plot cannot accept multiple colors when subplots=True (GH9894)
  • Bug in DataFrame.plot raises ValueError when color name is specified by multiple characters (GH10387)
  • Bug in left and right align of Series with MultiIndex may be inverted (GH10665)
  • Bug in left and right join of with MultiIndex may be inverted (GH10741)
  • Bug in read_stata when reading a file with a different order set in columns (GH10757)
  • Bug in Categorical may not representing properly when category contains tz or Period (GH10713)
  • Bug in Categorical.__iter__ may not returning correct datetime and Period (GH10713)
  • Bug in indexing with a PeriodIndex on an object with a PeriodIndex (GH4125)
  • Bug in read_csv with engine='c': EOF preceded by a comment, blank line, etc. was not handled correctly (GH10728, GH10548)
  • Reading “famafrench” data via DataReader results in HTTP 404 error because of the website url is changed (GH10591).
  • Bug in read_msgpack where DataFrame to decode has duplicate column names (GH9618)
  • Bug in io.common.get_filepath_or_buffer which caused reading of valid S3 files to fail if the bucket also contained keys for which the user does not have read permission (GH10604)
  • Bug in vectorised setting of timestamp columns with python datetime.date and numpy datetime64 (GH10408, GH10412)
  • Bug in Index.take may add unnecessary freq attribute (GH10791)
  • Bug in merge with empty DataFrame may raise IndexError (GH10824)
  • Bug in to_latex where unexpected keyword argument for some documented arguments (GH10888)
  • Bug in indexing of large DataFrame where IndexError is uncaught (GH10645 and GH10692)
  • Bug in read_csv when using the nrows or chunksize parameters if file contains only a header line (GH9535)
  • Bug in serialization of category types in HDF5 in presence of alternate encodings. (GH10366)
  • Bug in pd.DataFrame when constructing an empty DataFrame with a string dtype (GH9428)
  • Bug in pd.DataFrame.diff when DataFrame is not consolidated (GH10907)
  • Bug in pd.unique for arrays with the datetime64 or timedelta64 dtype that meant an array with object dtype was returned instead the original dtype (GH9431)
  • Bug in Timedelta raising error when slicing from 0s (GH10583)
  • Bug in DatetimeIndex.take and TimedeltaIndex.take may not raise IndexError against invalid index (GH10295)
  • Bug in Series([np.nan]).astype('M8[ms]'), which now returns Series([pd.NaT]) (GH10747)
  • Bug in PeriodIndex.order reset freq (GH10295)
  • Bug in date_range when freq divides end as nanos (GH10885)
  • Bug in iloc allowing memory outside bounds of a Series to be accessed with negative integers (GH10779)
  • Bug in read_msgpack where encoding is not respected (GH10581)
  • Bug preventing access to the first index when using iloc with a list containing the appropriate negative integer (GH10547, GH10779)
  • Bug in TimedeltaIndex formatter causing error while trying to save DataFrame with TimedeltaIndex using to_csv (GH10833)
  • Bug in DataFrame.where when handling Series slicing (GH10218, GH9558)
  • Bug where pd.read_gbq throws ValueError when Bigquery returns zero rows (GH10273)
  • Bug in to_json which was causing segmentation fault when serializing 0-rank ndarray (GH9576)
  • Bug in plotting functions may raise IndexError when plotted on GridSpec (GH10819)
  • Bug in plot result may show unnecessary minor ticklabels (GH10657)
  • Bug in groupby incorrect computation for aggregation on DataFrame with NaT (E.g first, last, min). (GH10590, GH11010)
  • Bug when constructing DataFrame where passing a dictionary with only scalar values and specifying columns did not raise an error (GH10856)
  • Bug in .var() causing roundoff errors for highly similar values (GH10242)
  • Bug in DataFrame.plot(subplots=True) with duplicated columns outputs incorrect result (GH10962)
  • Bug in Index arithmetic may result in incorrect class (GH10638)
  • Bug in date_range results in empty if freq is negative annualy, quarterly and monthly (GH11018)
  • Bug in DatetimeIndex cannot infer negative freq (GH11018)
  • Remove use of some deprecated numpy comparison operations, mainly in tests. (GH10569)
  • Bug in Index dtype may not applied properly (GH11017)
  • Bug in io.gbq when testing for minimum google api client version (GH10652)
  • Bug in DataFrame construction from nested dict with timedelta keys (GH11129)
  • Bug in .fillna against may raise TypeError when data contains datetime dtype (GH7095, GH11153)
  • Bug in .groupby when number of keys to group by is same as length of index (GH11185)
  • Bug in convert_objects where converted values might not be returned if all null and coerce (GH9589)
  • Bug in convert_objects where copy keyword was not respected (GH9589)

v0.16.2 (June 12, 2015)

This is a minor bug-fix release from 0.16.1 and includes a a large number of bug fixes along some new features (pipe() method), enhancements, and performance improvements.

We recommend that all users upgrade to this version.

Highlights include:

  • A new pipe method, see here
  • Documentation on how to use numba with pandas, see here

New features

Pipe

We’ve introduced a new method DataFrame.pipe(). As suggested by the name, pipe should be used to pipe data through a chain of function calls. The goal is to avoid confusing nested function calls like

# df is a DataFrame
# f, g, and h are functions that take and return DataFrames
f(g(h(df), arg1=1), arg2=2, arg3=3)

The logic flows from inside out, and function names are separated from their keyword arguments. This can be rewritten as

(df.pipe(h)
   .pipe(g, arg1=1)
   .pipe(f, arg2=2, arg3=3)
)

Now both the code and the logic flow from top to bottom. Keyword arguments are next to their functions. Overall the code is much more readable.

In the example above, the functions f, g, and h each expected the DataFrame as the first positional argument. When the function you wish to apply takes its data anywhere other than the first argument, pass a tuple of (function, keyword) indicating where the DataFrame should flow. For example:

In [1]: import statsmodels.formula.api as sm

In [2]: bb = pd.read_csv('data/baseball.csv', index_col='id')

# sm.poisson takes (formula, data)
In [3]: (bb.query('h > 0')
   ...:    .assign(ln_h = lambda df: np.log(df.h))
   ...:    .pipe((sm.poisson, 'data'), 'hr ~ ln_h + year + g + C(lg)')
   ...:    .fit()
   ...:    .summary()
   ...: )
   ...: 
Optimization terminated successfully.
         Current function value: 2.116284
         Iterations 24
Out[3]: 
<class 'statsmodels.iolib.summary.Summary'>
"""
                          Poisson Regression Results                          
==============================================================================
Dep. Variable:                     hr   No. Observations:                   68
Model:                        Poisson   Df Residuals:                       63
Method:                           MLE   Df Model:                            4
Date:                Don, 03 Nov 2016   Pseudo R-squ.:                  0.6878
Time:                        16:51:07   Log-Likelihood:                -143.91
converged:                       True   LL-Null:                       -460.91
                                        LLR p-value:                6.774e-136
===============================================================================
                  coef    std err          z      P>|z|      [95.0% Conf. Int.]
-------------------------------------------------------------------------------
Intercept   -1267.3636    457.867     -2.768      0.006     -2164.767  -369.960
C(lg)[T.NL]    -0.2057      0.101     -2.044      0.041        -0.403    -0.008
ln_h            0.9280      0.191      4.866      0.000         0.554     1.302
year            0.6301      0.228      2.762      0.006         0.183     1.077
g               0.0099      0.004      2.754      0.006         0.003     0.017
===============================================================================
"""

The pipe method is inspired by unix pipes, which stream text through processes. More recently dplyr and magrittr have introduced the popular (%>%) pipe operator for R.

See the documentation for more. (GH10129)

Other Enhancements

  • Added rsplit to Index/Series StringMethods (GH10303)

  • Removed the hard-coded size limits on the DataFrame HTML representation in the IPython notebook, and leave this to IPython itself (only for IPython v3.0 or greater). This eliminates the duplicate scroll bars that appeared in the notebook with large frames (GH10231).

    Note that the notebook has a toggle output scrolling feature to limit the display of very large frames (by clicking left of the output). You can also configure the way DataFrames are displayed using the pandas options, see here here.

  • axis parameter of DataFrame.quantile now accepts also index and column. (GH9543)

API Changes

  • Holiday now raises NotImplementedError if both offset and observance are used in the constructor instead of returning an incorrect result (GH10217).

Performance Improvements

  • Improved Series.resample performance with dtype=datetime64[ns] (GH7754)
  • Increase performance of str.split when expand=True (GH10081)

Bug Fixes

  • Bug in Series.hist raises an error when a one row Series was given (GH10214)
  • Bug where HDFStore.select modifies the passed columns list (GH7212)
  • Bug in Categorical repr with display.width of None in Python 3 (GH10087)
  • Bug in to_json with certain orients and a CategoricalIndex would segfault (GH10317)
  • Bug where some of the nan funcs do not have consistent return dtypes (GH10251)
  • Bug in DataFrame.quantile on checking that a valid axis was passed (GH9543)
  • Bug in groupby.apply aggregation for Categorical not preserving categories (GH10138)
  • Bug in to_csv where date_format is ignored if the datetime is fractional (GH10209)
  • Bug in DataFrame.to_json with mixed data types (GH10289)
  • Bug in cache updating when consolidating (GH10264)
  • Bug in mean() where integer dtypes can overflow (GH10172)
  • Bug where Panel.from_dict does not set dtype when specified (GH10058)
  • Bug in Index.union raises AttributeError when passing array-likes. (GH10149)
  • Bug in Timestamp‘s’ microsecond, quarter, dayofyear, week and daysinmonth properties return np.int type, not built-in int. (GH10050)
  • Bug in NaT raises AttributeError when accessing to daysinmonth, dayofweek properties. (GH10096)
  • Bug in Index repr when using the max_seq_items=None setting (GH10182).
  • Bug in getting timezone data with dateutil on various platforms ( GH9059, GH8639, GH9663, GH10121)
  • Bug in displaying datetimes with mixed frequencies; display ‘ms’ datetimes to the proper precision. (GH10170)
  • Bug in setitem where type promotion is applied to the entire block (GH10280)
  • Bug in Series arithmetic methods may incorrectly hold names (GH10068)
  • Bug in GroupBy.get_group when grouping on multiple keys, one of which is categorical. (GH10132)
  • Bug in DatetimeIndex and TimedeltaIndex names are lost after timedelta arithmetics ( GH9926)
  • Bug in DataFrame construction from nested dict with datetime64 (GH10160)
  • Bug in Series construction from dict with datetime64 keys (GH9456)
  • Bug in Series.plot(label="LABEL") not correctly setting the label (GH10119)
  • Bug in plot not defaulting to matplotlib axes.grid setting (GH9792)
  • Bug causing strings containing an exponent, but no decimal to be parsed as int instead of float in engine='python' for the read_csv parser (GH9565)
  • Bug in Series.align resets name when fill_value is specified (GH10067)
  • Bug in read_csv causing index name not to be set on an empty DataFrame (GH10184)
  • Bug in SparseSeries.abs resets name (GH10241)
  • Bug in TimedeltaIndex slicing may reset freq (GH10292)
  • Bug in GroupBy.get_group raises ValueError when group key contains NaT (GH6992)
  • Bug in SparseSeries constructor ignores input data name (GH10258)
  • Bug in Categorical.remove_categories causing a ValueError when removing the NaN category if underlying dtype is floating-point (GH10156)
  • Bug where infer_freq infers timerule (WOM-5XXX) unsupported by to_offset (GH9425)
  • Bug in DataFrame.to_hdf() where table format would raise a seemingly unrelated error for invalid (non-string) column names. This is now explicitly forbidden. (GH9057)
  • Bug to handle masking empty DataFrame (GH10126).
  • Bug where MySQL interface could not handle numeric table/column names (GH10255)
  • Bug in read_csv with a date_parser that returned a datetime64 array of other time resolution than [ns] (GH10245)
  • Bug in Panel.apply when the result has ndim=0 (GH10332)
  • Bug in read_hdf where auto_close could not be passed (GH9327).
  • Bug in read_hdf where open stores could not be used (GH10330).
  • Bug in adding empty DataFrame``s, now results in a ``DataFrame that .equals an empty DataFrame (GH10181).
  • Bug in to_hdf and HDFStore which did not check that complib choices were valid (GH4582, GH8874).

v0.16.1 (May 11, 2015)

This is a minor bug-fix release from 0.16.0 and includes a a large number of bug fixes along several new features, enhancements, and performance improvements. We recommend that all users upgrade to this version.

Highlights include:

  • Support for a CategoricalIndex, a category based index, see here
  • New section on how-to-contribute to pandas, see here
  • Revised “Merge, join, and concatenate” documentation, including graphical examples to make it easier to understand each operations, see here
  • New method sample for drawing random samples from Series, DataFrames and Panels. See here
  • The default Index printing has changed to a more uniform format, see here
  • BusinessHour datetime-offset is now supported, see here
  • Further enhancement to the .str accessor to make string operations easier, see here

Warning

In pandas 0.17.0, the sub-package pandas.io.data will be removed in favor of a separately installable package. See here for details (GH8961)

Enhancements

CategoricalIndex

We introduce a CategoricalIndex, a new type of index object that is useful for supporting indexing with duplicates. This is a container around a Categorical (introduced in v0.15.0) and allows efficient indexing and storage of an index with a large number of duplicated elements. Prior to 0.16.1, setting the index of a DataFrame/Series with a category dtype would convert this to regular object-based Index.

In [1]: df = DataFrame({'A' : np.arange(6),
   ...:                 'B' : Series(list('aabbca')).astype('category',
   ...:                                                     categories=list('cab'))
   ...:                })
   ...: 

In [2]: df
Out[2]: 
   A  B
0  0  a
1  1  a
2  2  b
3  3  b
4  4  c
5  5  a

In [3]: df.dtypes
Out[3]: 
A       int64
B    category
dtype: object

In [4]: df.B.cat.categories
Out[4]: Index([u'c', u'a', u'b'], dtype='object')

setting the index, will create create a CategoricalIndex

In [5]: df2 = df.set_index('B')

In [6]: df2.index
Out[6]: CategoricalIndex([u'a', u'a', u'b', u'b', u'c', u'a'], categories=[u'c', u'a', u'b'], ordered=False, name=u'B', dtype='category')

indexing with __getitem__/.iloc/.loc/.ix works similarly to an Index with duplicates. The indexers MUST be in the category or the operation will raise.

In [7]: df2.loc['a']
Out[7]: 
   A
B   
a  0
a  1
a  5

and preserves the CategoricalIndex

In [8]: df2.loc['a'].index
Out[8]: CategoricalIndex([u'a', u'a', u'a'], categories=[u'c', u'a', u'b'], ordered=False, name=u'B', dtype='category')

sorting will order by the order of the categories

In [9]: df2.sort_index()
Out[9]: 
   A
B   
c  4
a  0
a  1
a  5
b  2
b  3

groupby operations on the index will preserve the index nature as well

In [10]: df2.groupby(level=0).sum()
Out[10]: 
   A
B   
c  4
a  6
b  5

In [11]: df2.groupby(level=0).sum().index
Out[11]: CategoricalIndex([u'c', u'a', u'b'], categories=[u'c', u'a', u'b'], ordered=False, name=u'B', dtype='category')

reindexing operations, will return a resulting index based on the type of the passed indexer, meaning that passing a list will return a plain-old-Index; indexing with a Categorical will return a CategoricalIndex, indexed according to the categories of the PASSED Categorical dtype. This allows one to arbitrarly index these even with values NOT in the categories, similarly to how you can reindex ANY pandas index.

In [12]: df2.reindex(['a','e'])
Out[12]: 
     A
B     
a  0.0
a  1.0
a  5.0
e  NaN

In [13]: df2.reindex(['a','e']).index
Out[13]: Index([u'a', u'a', u'a', u'e'], dtype='object', name=u'B')

In [14]: df2.reindex(pd.Categorical(['a','e'],categories=list('abcde')))
Out[14]: 
     A
B     
a  0.0
a  1.0
a  5.0
e  NaN

In [15]: df2.reindex(pd.Categorical(['a','e'],categories=list('abcde'))).index
Out[15]: CategoricalIndex([u'a', u'a', u'a', u'e'], categories=[u'a', u'b', u'c', u'd', u'e'], ordered=False, name=u'B', dtype='category')

See the documentation for more. (GH7629, GH10038, GH10039)

Sample

Series, DataFrames, and Panels now have a new method: sample(). The method accepts a specific number of rows or columns to return, or a fraction of the total number or rows or columns. It also has options for sampling with or without replacement, for passing in a column for weights for non-uniform sampling, and for setting seed values to facilitate replication. (GH2419)

In [16]: example_series = Series([0,1,2,3,4,5])

# When no arguments are passed, returns 1
In [17]: example_series.sample()
Out[17]: 
3    3
dtype: int64

# One may specify either a number of rows:
In [18]: example_series.sample(n=3)
Out[18]: 
5    5
1    1
4    4
dtype: int64

# Or a fraction of the rows:
In [19]: example_series.sample(frac=0.5)
Out[19]: 
4    4
1    1
0    0
dtype: int64

# weights are accepted.
In [20]: example_weights = [0, 0, 0.2, 0.2, 0.2, 0.4]

In [21]: example_series.sample(n=3, weights=example_weights)
Out[21]: 
2    2
3    3
5    5
dtype: int64

# weights will also be normalized if they do not sum to one,
# and missing values will be treated as zeros.
In [22]: example_weights2 = [0.5, 0, 0, 0, None, np.nan]

In [23]: example_series.sample(n=1, weights=example_weights2)
Out[23]: 
0    0
dtype: int64

When applied to a DataFrame, one may pass the name of a column to specify sampling weights when sampling from rows.

In [24]: df = DataFrame({'col1':[9,8,7,6], 'weight_column':[0.5, 0.4, 0.1, 0]})

In [25]: df.sample(n=3, weights='weight_column')
Out[25]: 
   col1  weight_column
0     9            0.5
1     8            0.4
2     7            0.1

String Methods Enhancements

Continuing from v0.16.0, the following enhancements make string operations easier and more consistent with standard python string operations.

  • Added StringMethods (.str accessor) to Index (GH9068)

    The .str accessor is now available for both Series and Index.

    In [26]: idx = Index([' jack', 'jill ', ' jesse ', 'frank'])
    
    In [27]: idx.str.strip()
    Out[27]: Index([u'jack', u'jill', u'jesse', u'frank'], dtype='object')
    

    One special case for the .str accessor on Index is that if a string method returns bool, the .str accessor will return a np.array instead of a boolean Index (GH8875). This enables the following expression to work naturally:

    In [28]: idx = Index(['a1', 'a2', 'b1', 'b2'])
    
    In [29]: s = Series(range(4), index=idx)
    
    In [30]: s
    Out[30]: 
    a1    0
    a2    1
    b1    2
    b2    3
    dtype: int64
    
    In [31]: idx.str.startswith('a')
    Out[31]: array([ True,  True, False, False], dtype=bool)
    
    In [32]: s[s.index.str.startswith('a')]
    Out[32]: 
    a1    0
    a2    1
    dtype: int64
    
  • The following new methods are accesible via .str accessor to apply the function to each values. (GH9766, GH9773, GH10031, GH10045, GH10052)

    Methods
    capitalize() swapcase() normalize() partition() rpartition()
    index() rindex() translate()    
  • split now takes expand keyword to specify whether to expand dimensionality. return_type is deprecated. (GH9847)

    In [33]: s = Series(['a,b', 'a,c', 'b,c'])
    
    # return Series
    In [34]: s.str.split(',')
    Out[34]: 
    0    [a, b]
    1    [a, c]
    2    [b, c]
    dtype: object
    
    # return DataFrame
    In [35]: s.str.split(',', expand=True)
    Out[35]: 
       0  1
    0  a  b
    1  a  c
    2  b  c
    
    In [36]: idx = Index(['a,b', 'a,c', 'b,c'])
    
    # return Index
    In [37]: idx.str.split(',')
    Out[37]: Index([[u'a', u'b'], [u'a', u'c'], [u'b', u'c']], dtype='object')
    
    # return MultiIndex
    In [38]: idx.str.split(',', expand=True)
    Out[38]: 
    MultiIndex(levels=[[u'a', u'b'], [u'b', u'c']],
               labels=[[0, 0, 1], [0, 1, 1]])
    
  • Improved extract and get_dummies methods for Index.str (GH9980)

Other Enhancements

  • BusinessHour offset is now supported, which represents business hours starting from 09:00 - 17:00 on BusinessDay by default. See Here for details. (GH7905)

    In [39]: from pandas.tseries.offsets import BusinessHour
    
    In [40]: Timestamp('2014-08-01 09:00') + BusinessHour()
    Out[40]: Timestamp('2014-08-01 10:00:00')
    
    In [41]: Timestamp('2014-08-01 07:00') + BusinessHour()
    Out[41]: Timestamp('2014-08-01 10:00:00')
    
    In [42]: Timestamp('2014-08-01 16:30') + BusinessHour()
    Out[42]: Timestamp('2014-08-04 09:30:00')
    
  • DataFrame.diff now takes an axis parameter that determines the direction of differencing (GH9727)

  • Allow clip, clip_lower, and clip_upper to accept array-like arguments as thresholds (This is a regression from 0.11.0). These methods now have an axis parameter which determines how the Series or DataFrame will be aligned with the threshold(s). (GH6966)

  • DataFrame.mask() and Series.mask() now support same keywords as where (GH8801)

  • drop function can now accept errors keyword to suppress ValueError raised when any of label does not exist in the target data. (GH6736)

    In [43]: df = DataFrame(np.random.randn(3, 3), columns=['A', 'B', 'C'])
    
    In [44]: df.drop(['A', 'X'], axis=1, errors='ignore')
    Out[44]: 
              B         C
    0  1.058969 -0.397840
    1  1.047579  1.045938
    2 -0.122092  0.124713
    
  • Add support for separating years and quarters using dashes, for example 2014-Q1. (GH9688)

  • Allow conversion of values with dtype datetime64 or timedelta64 to strings using astype(str) (GH9757)

  • get_dummies function now accepts sparse keyword. If set to True, the return DataFrame is sparse, e.g. SparseDataFrame. (GH8823)

  • Period now accepts datetime64 as value input. (GH9054)

  • Allow timedelta string conversion when leading zero is missing from time definition, ie 0:00:00 vs 00:00:00. (GH9570)

  • Allow Panel.shift with axis='items' (GH9890)

  • Trying to write an excel file now raises NotImplementedError if the DataFrame has a MultiIndex instead of writing a broken Excel file. (GH9794)

  • Allow Categorical.add_categories to accept Series or np.array. (GH9927)

  • Add/delete str/dt/cat accessors dynamically from __dir__. (GH9910)

  • Add normalize as a dt accessor method. (GH10047)

  • DataFrame and Series now have _constructor_expanddim property as overridable constructor for one higher dimensionality data. This should be used only when it is really needed, see here

  • pd.lib.infer_dtype now returns 'bytes' in Python 3 where appropriate. (GH10032)

API changes

  • When passing in an ax to df.plot( ..., ax=ax), the sharex kwarg will now default to False. The result is that the visibility of xlabels and xticklabels will not anymore be changed. You have to do that by yourself for the right axes in your figure or set sharex=True explicitly (but this changes the visible for all axes in the figure, not only the one which is passed in!). If pandas creates the subplots itself (e.g. no passed in ax kwarg), then the default is still sharex=True and the visibility changes are applied.
  • assign() now inserts new columns in alphabetical order. Previously the order was arbitrary. (GH9777)
  • By default, read_csv and read_table will now try to infer the compression type based on the file extension. Set compression=None to restore the previous behavior (no decompression). (GH9770)

Deprecations

  • Series.str.split‘s return_type keyword was removed in favor of expand (GH9847)

Index Representation

The string representation of Index and its sub-classes have now been unified. These will show a single-line display if there are few values; a wrapped multi-line display for a lot of values (but less than display.max_seq_items; if lots of items (> display.max_seq_items) will show a truncated display (the head and tail of the data). The formatting for MultiIndex is unchanges (a multi-line wrapped display). The display width responds to the option display.max_seq_items, which is defaulted to 100. (GH6482)

Previous Behavior

In [2]: pd.Index(range(4),name='foo')
Out[2]: Int64Index([0, 1, 2, 3], dtype='int64')

In [3]: pd.Index(range(104),name='foo')
Out[3]: Int64Index([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, ...], dtype='int64')

In [4]: pd.date_range('20130101',periods=4,name='foo',tz='US/Eastern')
Out[4]:
<class 'pandas.tseries.index.DatetimeIndex'>
[2013-01-01 00:00:00-05:00, ..., 2013-01-04 00:00:00-05:00]
Length: 4, Freq: D, Timezone: US/Eastern

In [5]: pd.date_range('20130101',periods=104,name='foo',tz='US/Eastern')
Out[5]:
<class 'pandas.tseries.index.DatetimeIndex'>
[2013-01-01 00:00:00-05:00, ..., 2013-04-14 00:00:00-04:00]
Length: 104, Freq: D, Timezone: US/Eastern

New Behavior

In [45]: pd.set_option('display.width', 80)

In [46]: pd.Index(range(4), name='foo')
Out[46]: Int64Index([0, 1, 2, 3], dtype='int64', name=u'foo')

In [47]: pd.Index(range(30), name='foo')
Out[47]: 
Int64Index([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,
            17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29],
           dtype='int64', name=u'foo')

In [48]: pd.Index(range(104), name='foo')
Out[48]: 
Int64Index([  0,   1,   2,   3,   4,   5,   6,   7,   8,   9,
            ...
             94,  95,  96,  97,  98,  99, 100, 101, 102, 103],
           dtype='int64', name=u'foo', length=104)

In [49]: pd.CategoricalIndex(['a','bb','ccc','dddd'], ordered=True, name='foobar')
Out[49]: CategoricalIndex([u'a', u'bb', u'ccc', u'dddd'], categories=[u'a', u'bb', u'ccc', u'dddd'], ordered=True, name=u'foobar', dtype='category')

In [50]: pd.CategoricalIndex(['a','bb','ccc','dddd']*10, ordered=True, name='foobar')
Out[50]: 
CategoricalIndex([u'a', u'bb', u'ccc', u'dddd', u'a', u'bb', u'ccc', u'dddd',
                  u'a', u'bb', u'ccc', u'dddd', u'a', u'bb', u'ccc', u'dddd',
                  u'a', u'bb', u'ccc', u'dddd', u'a', u'bb', u'ccc', u'dddd',
                  u'a', u'bb', u'ccc', u'dddd', u'a', u'bb', u'ccc', u'dddd',
                  u'a', u'bb', u'ccc', u'dddd', u'a', u'bb', u'ccc', u'dddd'],
                 categories=[u'a', u'bb', u'ccc', u'dddd'], ordered=True, name=u'foobar', dtype='category')

In [51]: pd.CategoricalIndex(['a','bb','ccc','dddd']*100, ordered=True, name='foobar')
Out[51]: 
CategoricalIndex([u'a', u'bb', u'ccc', u'dddd', u'a', u'bb', u'ccc', u'dddd',
                  u'a', u'bb',
                  ...
                  u'ccc', u'dddd', u'a', u'bb', u'ccc', u'dddd', u'a', u'bb',
                  u'ccc', u'dddd'],
                 categories=[u'a', u'bb', u'ccc', u'dddd'], ordered=True, name=u'foobar', dtype='category', length=400)

In [52]: pd.date_range('20130101',periods=4, name='foo', tz='US/Eastern')
Out[52]: 
DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 00:00:00-05:00',
               '2013-01-03 00:00:00-05:00', '2013-01-04 00:00:00-05:00'],
              dtype='datetime64[ns, US/Eastern]', name=u'foo', freq='D')

In [53]: pd.date_range('20130101',periods=25, freq='D')
Out[53]: 
DatetimeIndex(['2013-01-01', '2013-01-02', '2013-01-03', '2013-01-04',
               '2013-01-05', '2013-01-06', '2013-01-07', '2013-01-08',
               '2013-01-09', '2013-01-10', '2013-01-11', '2013-01-12',
               '2013-01-13', '2013-01-14', '2013-01-15', '2013-01-16',
               '2013-01-17', '2013-01-18', '2013-01-19', '2013-01-20',
               '2013-01-21', '2013-01-22', '2013-01-23', '2013-01-24',
               '2013-01-25'],
              dtype='datetime64[ns]', freq='D')

In [54]: pd.date_range('20130101',periods=104, name='foo', tz='US/Eastern')
Out[54]: 
DatetimeIndex(['2013-01-01 00:00:00-05:00', '2013-01-02 00:00:00-05:00',
               '2013-01-03 00:00:00-05:00', '2013-01-04 00:00:00-05:00',
               '2013-01-05 00:00:00-05:00', '2013-01-06 00:00:00-05:00',
               '2013-01-07 00:00:00-05:00', '2013-01-08 00:00:00-05:00',
               '2013-01-09 00:00:00-05:00', '2013-01-10 00:00:00-05:00',
               ...
               '2013-04-05 00:00:00-04:00', '2013-04-06 00:00:00-04:00',
               '2013-04-07 00:00:00-04:00', '2013-04-08 00:00:00-04:00',
               '2013-04-09 00:00:00-04:00', '2013-04-10 00:00:00-04:00',
               '2013-04-11 00:00:00-04:00', '2013-04-12 00:00:00-04:00',
               '2013-04-13 00:00:00-04:00', '2013-04-14 00:00:00-04:00'],
              dtype='datetime64[ns, US/Eastern]', name=u'foo', length=104, freq='D')

Performance Improvements

  • Improved csv write performance with mixed dtypes, including datetimes by up to 5x (GH9940)
  • Improved csv write performance generally by 2x (GH9940)
  • Improved the performance of pd.lib.max_len_string_array by 5-7x (GH10024)

Bug Fixes

  • Bug where labels did not appear properly in the legend of DataFrame.plot(), passing label= arguments works, and Series indices are no longer mutated. (GH9542)
  • Bug in json serialization causing a segfault when a frame had zero length. (GH9805)
  • Bug in read_csv where missing trailing delimiters would cause segfault. (GH5664)
  • Bug in retaining index name on appending (GH9862)
  • Bug in scatter_matrix draws unexpected axis ticklabels (GH5662)
  • Fixed bug in StataWriter resulting in changes to input DataFrame upon save (GH9795).
  • Bug in transform causing length mismatch when null entries were present and a fast aggregator was being used (GH9697)
  • Bug in equals causing false negatives when block order differed (GH9330)
  • Bug in grouping with multiple pd.Grouper where one is non-time based (GH10063)
  • Bug in read_sql_table error when reading postgres table with timezone (GH7139)
  • Bug in DataFrame slicing may not retain metadata (GH9776)
  • Bug where TimdeltaIndex were not properly serialized in fixed HDFStore (GH9635)
  • Bug with TimedeltaIndex constructor ignoring name when given another TimedeltaIndex as data (GH10025).
  • Bug in DataFrameFormatter._get_formatted_index with not applying max_colwidth to the DataFrame index (GH7856)
  • Bug in .loc with a read-only ndarray data source (GH10043)
  • Bug in groupby.apply() that would raise if a passed user defined function either returned only None (for all input). (GH9685)
  • Always use temporary files in pytables tests (GH9992)
  • Bug in plotting continuously using secondary_y may not show legend properly. (GH9610, GH9779)
  • Bug in DataFrame.plot(kind="hist") results in TypeError when DataFrame contains non-numeric columns (GH9853)
  • Bug where repeated plotting of DataFrame with a DatetimeIndex may raise TypeError (GH9852)
  • Bug in setup.py that would allow an incompat cython version to build (GH9827)
  • Bug in plotting secondary_y incorrectly attaches right_ax property to secondary axes specifying itself recursively. (GH9861)
  • Bug in Series.quantile on empty Series of type Datetime or Timedelta (GH9675)
  • Bug in where causing incorrect results when upcasting was required (GH9731)
  • Bug in FloatArrayFormatter where decision boundary for displaying “small” floats in decimal format is off by one order of magnitude for a given display.precision (GH9764)
  • Fixed bug where DataFrame.plot() raised an error when both color and style keywords were passed and there was no color symbol in the style strings (GH9671)
  • Not showing a DeprecationWarning on combining list-likes with an Index (GH10083)
  • Bug in read_csv and read_table when using skip_rows parameter if blank lines are present. (GH9832)
  • Bug in read_csv() interprets index_col=True as 1 (GH9798)
  • Bug in index equality comparisons using == failing on Index/MultiIndex type incompatibility (GH9785)
  • Bug in which SparseDataFrame could not take nan as a column name (GH8822)
  • Bug in to_msgpack and read_msgpack zlib and blosc compression support (GH9783)
  • Bug GroupBy.size doesn’t attach index name properly if grouped by TimeGrouper (GH9925)
  • Bug causing an exception in slice assignments because length_of_indexer returns wrong results (GH9995)
  • Bug in csv parser causing lines with initial whitespace plus one non-space character to be skipped. (GH9710)
  • Bug in C csv parser causing spurious NaNs when data started with newline followed by whitespace. (GH10022)
  • Bug causing elements with a null group to spill into the final group when grouping by a Categorical (GH9603)
  • Bug where .iloc and .loc behavior is not consistent on empty dataframes (GH9964)
  • Bug in invalid attribute access on a TimedeltaIndex incorrectly raised ValueError instead of AttributeError (GH9680)
  • Bug in unequal comparisons between categorical data and a scalar, which was not in the categories (e.g. Series(Categorical(list("abc"), ordered=True)) > "d". This returned False for all elements, but now raises a TypeError. Equality comparisons also now return False for == and True for !=. (GH9848)
  • Bug in DataFrame __setitem__ when right hand side is a dictionary (GH9874)
  • Bug in where when dtype is datetime64/timedelta64, but dtype of other is not (GH9804)
  • Bug in MultiIndex.sortlevel() results in unicode level name breaks (GH9856)
  • Bug in which groupby.transform incorrectly enforced output dtypes to match input dtypes. (GH9807)
  • Bug in DataFrame constructor when columns parameter is set, and data is an empty list (GH9939)
  • Bug in bar plot with log=True raises TypeError if all values are less than 1 (GH9905)
  • Bug in horizontal bar plot ignores log=True (GH9905)
  • Bug in PyTables queries that did not return proper results using the index (GH8265, GH9676)
  • Bug where dividing a dataframe containing values of type Decimal by another Decimal would raise. (GH9787)
  • Bug where using DataFrames asfreq would remove the name of the index. (GH9885)
  • Bug causing extra index point when resample BM/BQ (GH9756)
  • Changed caching in AbstractHolidayCalendar to be at the instance level rather than at the class level as the latter can result in unexpected behaviour. (GH9552)
  • Fixed latex output for multi-indexed dataframes (GH9778)
  • Bug causing an exception when setting an empty range using DataFrame.loc (GH9596)
  • Bug in hiding ticklabels with subplots and shared axes when adding a new plot to an existing grid of axes (GH9158)
  • Bug in transform and filter when grouping on a categorical variable (GH9921)
  • Bug in transform when groups are equal in number and dtype to the input index (GH9700)
  • Google BigQuery connector now imports dependencies on a per-method basis.(GH9713)
  • Updated BigQuery connector to no longer use deprecated oauth2client.tools.run() (GH8327)
  • Bug in subclassed DataFrame. It may not return the correct class, when slicing or subsetting it. (GH9632)
  • Bug in .median() where non-float null values are not handled correctly (GH10040)
  • Bug in Series.fillna() where it raises if a numerically convertible string is given (GH10092)

v0.16.0 (March 22, 2015)

This is a major release from 0.15.2 and includes a small number of API changes, several new features, enhancements, and performance improvements along with a large number of bug fixes. We recommend that all users upgrade to this version.

Highlights include:

  • DataFrame.assign method, see here
  • Series.to_coo/from_coo methods to interact with scipy.sparse, see here
  • Backwards incompatible change to Timedelta to conform the .seconds attribute with datetime.timedelta, see here
  • Changes to the .loc slicing API to conform with the behavior of .ix see here
  • Changes to the default for ordering in the Categorical constructor, see here
  • Enhancement to the .str accessor to make string operations easier, see here
  • The pandas.tools.rplot, pandas.sandbox.qtpandas and pandas.rpy modules are deprecated. We refer users to external packages like seaborn, pandas-qt and rpy2 for similar or equivalent functionality, see here

Check the API Changes and deprecations before updating.

New features

DataFrame Assign

Inspired by dplyr’s mutate verb, DataFrame has a new assign() method. The function signature for assign is simply **kwargs. The keys are the column names for the new fields, and the values are either a value to be inserted (for example, a Series or NumPy array), or a function of one argument to be called on the DataFrame. The new values are inserted, and the entire DataFrame (with all original and new columns) is returned.

In [1]: iris = read_csv('data/iris.data')

In [2]: iris.head()
Out[2]: 
   SepalLength  SepalWidth  PetalLength  PetalWidth         Name
0          5.1         3.5          1.4         0.2  Iris-setosa
1          4.9         3.0          1.4         0.2  Iris-setosa
2          4.7         3.2          1.3         0.2  Iris-setosa
3          4.6         3.1          1.5         0.2  Iris-setosa
4          5.0         3.6          1.4         0.2  Iris-setosa

In [3]: iris.assign(sepal_ratio=iris['SepalWidth'] / iris['SepalLength']).head()
Out[3]: 
   SepalLength  SepalWidth  PetalLength  PetalWidth         Name  sepal_ratio
0          5.1         3.5          1.4         0.2  Iris-setosa     0.686275
1          4.9         3.0          1.4         0.2  Iris-setosa     0.612245
2          4.7         3.2          1.3         0.2  Iris-setosa     0.680851
3          4.6         3.1          1.5         0.2  Iris-setosa     0.673913
4          5.0         3.6          1.4         0.2  Iris-setosa     0.720000

Above was an example of inserting a precomputed value. We can also pass in a function to be evalutated.

In [4]: iris.assign(sepal_ratio = lambda x: (x['SepalWidth'] /
   ...:                                      x['SepalLength'])).head()
   ...: 
Out[4]: 
   SepalLength  SepalWidth  PetalLength  PetalWidth         Name  sepal_ratio
0          5.1         3.5          1.4         0.2  Iris-setosa     0.686275
1          4.9         3.0          1.4         0.2  Iris-setosa     0.612245
2          4.7         3.2          1.3         0.2  Iris-setosa     0.680851
3          4.6         3.1          1.5         0.2  Iris-setosa     0.673913
4          5.0         3.6          1.4         0.2  Iris-setosa     0.720000

The power of assign comes when used in chains of operations. For example, we can limit the DataFrame to just those with a Sepal Length greater than 5, calculate the ratio, and plot

In [5]: (iris.query('SepalLength > 5')
   ...:      .assign(SepalRatio = lambda x: x.SepalWidth / x.SepalLength,
   ...:              PetalRatio = lambda x: x.PetalWidth / x.PetalLength)
   ...:      .plot(kind='scatter', x='SepalRatio', y='PetalRatio'))
   ...: 
Out[5]: <matplotlib.axes._subplots.AxesSubplot at 0x7fd23589bb10>
_images/whatsnew_assign.png

See the documentation for more. (GH9229)

Interaction with scipy.sparse

Added SparseSeries.to_coo() and SparseSeries.from_coo() methods (GH8048) for converting to and from scipy.sparse.coo_matrix instances (see here). For example, given a SparseSeries with MultiIndex we can convert to a scipy.sparse.coo_matrix by specifying the row and column labels as index levels:

In [6]: from numpy import nan

In [7]: s = Series([3.0, nan, 1.0, 3.0, nan, nan])

In [8]: s.index = MultiIndex.from_tuples([(1, 2, 'a', 0),
   ...:                                   (1, 2, 'a', 1),
   ...:                                   (1, 1, 'b', 0),
   ...:                                   (1, 1, 'b', 1),
   ...:                                   (2, 1, 'b', 0),
   ...:                                   (2, 1, 'b', 1)],
   ...:                                   names=['A', 'B', 'C', 'D'])
   ...: 

In [9]: s
Out[9]: 
A  B  C  D
1  2  a  0    3.0
         1    NaN
   1  b  0    1.0
         1    3.0
2  1  b  0    NaN
         1    NaN
dtype: float64

# SparseSeries
In [10]: ss = s.to_sparse()

In [11]: ss
Out[11]: 
A  B  C  D
1  2  a  0    3.0
         1    NaN
   1  b  0    1.0
         1    3.0
2  1  b  0    NaN
         1    NaN
dtype: float64
BlockIndex
Block locations: array([0, 2], dtype=int32)
Block lengths: array([1, 2], dtype=int32)

In [12]: A, rows, columns = ss.to_coo(row_levels=['A', 'B'],
   ....:                              column_levels=['C', 'D'],
   ....:                              sort_labels=False)
   ....: 

In [13]: A
Out[13]: 
<3x4 sparse matrix of type '<type 'numpy.float64'>'
	with 3 stored elements in COOrdinate format>

In [14]: A.todense()
Out[14]: 
matrix([[ 3.,  0.,  0.,  0.],
        [ 0.,  0.,  1.,  3.],
        [ 0.,  0.,  0.,  0.]])

In [15]: rows
Out[15]: [(1, 2), (1, 1), (2, 1)]

In [16]: columns
Out[16]: [('a', 0), ('a', 1), ('b', 0), ('b', 1)]

The from_coo method is a convenience method for creating a SparseSeries from a scipy.sparse.coo_matrix:

In [17]: from scipy import sparse

In [18]: A = sparse.coo_matrix(([3.0, 1.0, 2.0], ([1, 0, 0], [0, 2, 3])),
   ....:                             shape=(3, 4))
   ....: 

In [19]: A
Out[19]: 
<3x4 sparse matrix of type '<type 'numpy.float64'>'
	with 3 stored elements in COOrdinate format>

In [20]: A.todense()
Out[20]: 
matrix([[ 0.,  0.,  1.,  2.],
        [ 3.,  0.,  0.,  0.],
        [ 0.,  0.,  0.,  0.]])

In [21]: ss = SparseSeries.from_coo(A)

In [22]: ss
Out[22]: 
0  2    1.0
   3    2.0
1  0    3.0
dtype: float64
BlockIndex
Block locations: array([0], dtype=int32)
Block lengths: array([3], dtype=int32)

String Methods Enhancements

  • Following new methods are accesible via .str accessor to apply the function to each values. This is intended to make it more consistent with standard methods on strings. (GH9282, GH9352, GH9386, GH9387, GH9439)

    Methods
    isalnum() isalpha() isdigit() isdigit() isspace()
    islower() isupper() istitle() isnumeric() isdecimal()
    find() rfind()