Version 0.11.0 (April 22, 2013)¶
This is a major release from 0.10.1 and includes many new features and enhancements along with a large number of bug fixes. The methods of Selecting Data have had quite a number of additions, and Dtype support is now full-fledged. There are also a number of important API changes that long-time pandas users should pay close attention to.
There is a new section in the documentation, 10 Minutes to Pandas, primarily geared to new users.
There is a new section in the documentation, Cookbook, a collection of useful recipes in pandas (and that we want contributions!).
There are several libraries that are now Recommended Dependencies
Selection choices¶
Starting in 0.11.0, object selection has had a number of user-requested additions in order to support more explicit location based indexing. pandas now supports three types of multi-axis indexing.
.loc
is strictly label based, will raiseKeyError
when the items are not found, allowed inputs are:A single label, e.g.
5
or'a'
, (note that5
is interpreted as a label of the index. This use is not an integer position along the index)A list or array of labels
['a', 'b', 'c']
A slice object with labels
'a':'f'
, (note that contrary to usual python slices, both the start and the stop are included!)A boolean array
See more at Selection by Label
.iloc
is strictly integer position based (from0
tolength-1
of the axis), will raiseIndexError
when the requested indices are out of bounds. Allowed inputs are:An integer e.g.
5
A list or array of integers
[4, 3, 0]
A slice object with ints
1:7
A boolean array
See more at Selection by Position
.ix
supports mixed integer and label based access. It is primarily label based, but will fallback to integer positional access..ix
is the most general and will support any of the inputs to.loc
and.iloc
, as well as support for floating point label schemes..ix
is especially useful when dealing with mixed positional and label based hierarchical indexes.As using integer slices with
.ix
have different behavior depending on whether the slice is interpreted as position based or label based, it’s usually better to be explicit and use.iloc
or.loc
.See more at Advanced Indexing and Advanced Hierarchical.
Selection deprecations¶
Starting in version 0.11.0, these methods may be deprecated in future versions.
irow
icol
iget_value
See the section Selection by Position for substitutes.
Dtypes¶
Numeric dtypes will propagate and can coexist in DataFrames. If a dtype is passed (either directly via the dtype
keyword, a passed ndarray
, or a passed Series
, then it will be preserved in DataFrame operations. Furthermore, different numeric dtypes will NOT be combined. The following example will give you a taste.
In [1]: df1 = pd.DataFrame(np.random.randn(8, 1), columns=['A'], dtype='float32')
In [2]: df1
Out[2]:
A
0 0.469112
1 -0.282863
2 -1.509058
3 -1.135632
4 1.212112
5 -0.173215
6 0.119209
7 -1.044236
In [3]: df1.dtypes
Out[3]:
A float32
dtype: object
In [4]: df2 = pd.DataFrame({'A': pd.Series(np.random.randn(8), dtype='float16'),
...: 'B': pd.Series(np.random.randn(8)),
...: 'C': pd.Series(range(8), dtype='uint8')})
...:
In [5]: df2
Out[5]:
A B C
0 -0.861816 -0.424972 0
1 -2.105469 0.567020 1
2 -0.494873 0.276232 2
3 1.072266 -1.087401 3
4 0.721680 -0.673690 4
5 -0.706543 0.113648 5
6 -1.040039 -1.478427 6
7 0.271973 0.524988 7
In [6]: df2.dtypes
Out[6]:
A float16
B float64
C uint8
dtype: object
# here you get some upcasting
In [7]: df3 = df1.reindex_like(df2).fillna(value=0.0) + df2
In [8]: df3
Out[8]:
A B C
0 -0.392704 -0.424972 0.0
1 -2.388332 0.567020 1.0
2 -2.003932 0.276232 2.0
3 -0.063367 -1.087401 3.0
4 1.933792 -0.673690 4.0
5 -0.879758 0.113648 5.0
6 -0.920830 -1.478427 6.0
7 -0.772263 0.524988 7.0
In [9]: df3.dtypes
Out[9]:
A float32
B float64
C float64
dtype: object
Dtype conversion¶
This is lower-common-denominator upcasting, meaning you get the dtype which can accommodate all of the types
In [10]: df3.values.dtype
Out[10]: dtype('float64')
Conversion
In [11]: df3.astype('float32').dtypes
Out[11]:
A float32
B float32
C float32
dtype: object
Mixed conversion
In [12]: df3['D'] = '1.'
In [13]: df3['E'] = '1'
In [14]: df3.convert_objects(convert_numeric=True).dtypes
Out[14]:
A float32
B float64
C float64
D float64
E int64
dtype: object
# same, but specific dtype conversion
In [15]: df3['D'] = df3['D'].astype('float16')
In [16]: df3['E'] = df3['E'].astype('int32')
In [17]: df3.dtypes
Out[17]:
A float32
B float64
C float64
D float16
E int32
dtype: object
Forcing date coercion (and setting NaT
when not datelike)
In [18]: import datetime
In [19]: s = pd.Series([datetime.datetime(2001, 1, 1, 0, 0), 'foo', 1.0, 1,
....: pd.Timestamp('20010104'), '20010105'], dtype='O')
....:
In [20]: s.convert_objects(convert_dates='coerce')
Out[20]:
0 2001-01-01
1 NaT
2 NaT
3 NaT
4 2001-01-04
5 2001-01-05
dtype: datetime64[ns]
Dtype gotchas¶
Platform gotchas
Starting in 0.11.0, construction of DataFrame/Series will use default dtypes of int64
and float64
,
regardless of platform. This is not an apparent change from earlier versions of pandas. If you specify
dtypes, they WILL be respected, however (GH2837)
The following will all result in int64
dtypes
In [21]: pd.DataFrame([1, 2], columns=['a']).dtypes
Out[21]:
a int64
dtype: object
In [22]: pd.DataFrame({'a': [1, 2]}).dtypes
Out[22]:
a int64
dtype: object
In [23]: pd.DataFrame({'a': 1}, index=range(2)).dtypes
Out[23]:
a int64
dtype: object
Keep in mind that DataFrame(np.array([1,2]))
WILL result in int32
on 32-bit platforms!
Upcasting gotchas
Performing indexing operations on integer type data can easily upcast the data.
The dtype of the input data will be preserved in cases where nans
are not introduced.
In [24]: dfi = df3.astype('int32')
In [25]: dfi['D'] = dfi['D'].astype('int64')
In [26]: dfi
Out[26]:
A B C D E
0 0 0 0 1 1
1 -2 0 1 1 1
2 -2 0 2 1 1
3 0 -1 3 1 1
4 1 0 4 1 1
5 0 0 5 1 1
6 0 -1 6 1 1
7 0 0 7 1 1
In [27]: dfi.dtypes
Out[27]:
A int32
B int32
C int32
D int64
E int32
dtype: object
In [28]: casted = dfi[dfi > 0]
In [29]: casted
Out[29]:
A B C D E
0 NaN NaN NaN 1 1
1 NaN NaN 1.0 1 1
2 NaN NaN 2.0 1 1
3 NaN NaN 3.0 1 1
4 1.0 NaN 4.0 1 1
5 NaN NaN 5.0 1 1
6 NaN NaN 6.0 1 1
7 NaN NaN 7.0 1 1
In [30]: casted.dtypes
Out[30]:
A float64
B float64
C float64
D int64
E int32
dtype: object
While float dtypes are unchanged.
In [31]: df4 = df3.copy()
In [32]: df4['A'] = df4['A'].astype('float32')
In [33]: df4.dtypes
Out[33]:
A float32
B float64
C float64
D float16
E int32
dtype: object
In [34]: casted = df4[df4 > 0]
In [35]: casted
Out[35]:
A B C D E
0 NaN NaN NaN 1.0 1
1 NaN 0.567020 1.0 1.0 1
2 NaN 0.276232 2.0 1.0 1
3 NaN NaN 3.0 1.0 1
4 1.933792 NaN 4.0 1.0 1
5 NaN 0.113648 5.0 1.0 1
6 NaN NaN 6.0 1.0 1
7 NaN 0.524988 7.0 1.0 1
In [36]: casted.dtypes
Out[36]:
A float32
B float64
C float64
D float16
E int32
dtype: object
Datetimes conversion¶
Datetime64[ns] columns in a DataFrame (or a Series) allow the use of np.nan
to indicate a nan value,
in addition to the traditional NaT
, or not-a-time. This allows convenient nan setting in a generic way.
Furthermore datetime64[ns]
columns are created by default, when passed datetimelike objects (this change was introduced in 0.10.1)
(GH2809, GH2810)
In [12]: df = pd.DataFrame(np.random.randn(6, 2), pd.date_range('20010102', periods=6),
....: columns=['A', ' B'])
....:
In [13]: df['timestamp'] = pd.Timestamp('20010103')
In [14]: df
Out[14]:
A B timestamp
2001-01-02 0.404705 0.577046 2001-01-03
2001-01-03 -1.715002 -1.039268 2001-01-03
2001-01-04 -0.370647 -1.157892 2001-01-03
2001-01-05 -1.344312 0.844885 2001-01-03
2001-01-06 1.075770 -0.109050 2001-01-03
2001-01-07 1.643563 -1.469388 2001-01-03
# datetime64[ns] out of the box
In [15]: df.dtypes.value_counts()
Out[15]:
float64 2
datetime64[ns] 1
dtype: int64
# use the traditional nan, which is mapped to NaT internally
In [16]: df.loc[df.index[2:4], ['A', 'timestamp']] = np.nan
In [17]: df
Out[17]:
A B timestamp
2001-01-02 0.404705 0.577046 2001-01-03
2001-01-03 -1.715002 -1.039268 2001-01-03
2001-01-04 NaN -1.157892 NaT
2001-01-05 NaN 0.844885 NaT
2001-01-06 1.075770 -0.109050 2001-01-03
2001-01-07 1.643563 -1.469388 2001-01-03
Astype conversion on datetime64[ns]
to object
, implicitly converts NaT
to np.nan
In [18]: import datetime
In [19]: s = pd.Series([datetime.datetime(2001, 1, 2, 0, 0) for i in range(3)])
In [20]: s.dtype
Out[20]: dtype('<M8[ns]')
In [21]: s[1] = np.nan
In [22]: s
Out[22]:
0 2001-01-02
1 NaT
2 2001-01-02
dtype: datetime64[ns]
In [23]: s.dtype
Out[23]: dtype('<M8[ns]')
In [24]: s = s.astype('O')
In [25]: s
Out[25]:
0 2001-01-02 00:00:00
1 NaT
2 2001-01-02 00:00:00
dtype: object
In [26]: s.dtype
Out[26]: dtype('O')
API changes¶
Added to_series() method to indices, to facilitate the creation of indexers (GH3275)
HDFStore
added the method
select_column
to select a single column from a table as a Series.deprecated the
unique
method, can be replicated byselect_column(key,column).unique()
min_itemsize
parameter toappend
will now automatically create data_columns for passed keys
Enhancements¶
Improved performance of df.to_csv() by up to 10x in some cases. (GH3059)
Numexpr is now a Recommended Dependencies, to accelerate certain types of numerical and boolean operations
Bottleneck is now a Recommended Dependencies, to accelerate certain types of
nan
operations
HDFStore
support
read_hdf/to_hdf
API similar toread_csv/to_csv
In [27]: df = pd.DataFrame({'A': range(5), 'B': range(5)}) In [28]: df.to_hdf('store.h5', 'table', append=True) In [29]: pd.read_hdf('store.h5', 'table', where=['index > 2']) Out[29]: A B 3 3 3 4 4 4provide dotted attribute access to
get
from stores, e.g.store.df == store['df']
new keywords
iterator=boolean
, andchunksize=number_in_a_chunk
are provided to support iteration onselect
andselect_as_multiple
(GH3076)You can now select timestamps from an unordered timeseries similarly to an ordered timeseries (GH2437)
You can now select with a string from a DataFrame with a datelike index, in a similar way to a Series (GH3070)
In [30]: idx = pd.date_range("2001-10-1", periods=5, freq='M') In [31]: ts = pd.Series(np.random.rand(len(idx)), index=idx) In [32]: ts['2001'] Out[32]: 2001-10-31 0.117967 2001-11-30 0.702184 2001-12-31 0.414034 Freq: M, dtype: float64 In [33]: df = pd.DataFrame({'A': ts}) In [34]: df['2001'] Out[34]: A 2001-10-31 0.117967 2001-11-30 0.702184 2001-12-31 0.414034
Squeeze
to possibly remove length 1 dimensions from an object.>>> p = pd.Panel(np.random.randn(3, 4, 4), items=['ItemA', 'ItemB', 'ItemC'], ... major_axis=pd.date_range('20010102', periods=4), ... minor_axis=['A', 'B', 'C', 'D']) >>> p <class 'pandas.core.panel.Panel'> Dimensions: 3 (items) x 4 (major_axis) x 4 (minor_axis) Items axis: ItemA to ItemC Major_axis axis: 2001-01-02 00:00:00 to 2001-01-05 00:00:00 Minor_axis axis: A to D >>> p.reindex(items=['ItemA']).squeeze() A B C D 2001-01-02 0.926089 -2.026458 0.501277 -0.204683 2001-01-03 -0.076524 1.081161 1.141361 0.479243 2001-01-04 0.641817 -0.185352 1.824568 0.809152 2001-01-05 0.575237 0.669934 1.398014 -0.399338 >>> p.reindex(items=['ItemA'], minor=['B']).squeeze() 2001-01-02 -2.026458 2001-01-03 1.081161 2001-01-04 -0.185352 2001-01-05 0.669934 Freq: D, Name: B, dtype: float64In
pd.io.data.Options
,
Fix bug when trying to fetch data for the current month when already past expiry.
Now using lxml to scrape html instead of BeautifulSoup (lxml was faster).
New instance variables for calls and puts are automatically created when a method that creates them is called. This works for current month where the instance variables are simply
calls
andputs
. Also works for future expiry months and save the instance variable ascallsMMYY
orputsMMYY
, whereMMYY
are, respectively, the month and year of the option’s expiry.
Options.get_near_stock_price
now allows the user to specify the month for which to get relevant options data.
Options.get_forward_data
now has optional kwargsnear
andabove_below
. This allows the user to specify if they would like to only return forward looking data for options near the current stock price. This just obtains the data from Options.get_near_stock_price instead of Options.get_xxx_data() (GH2758).Cursor coordinate information is now displayed in time-series plots.
added option
display.max_seq_items
to control the number of elements printed per sequence pprinting it. (GH2979)added option
display.chop_threshold
to control display of small numerical values. (GH2739)added option
display.max_info_rows
to prevent verbose_info from being calculated for frames above 1M rows (configurable). (GH2807, GH2918)value_counts() now accepts a “normalize” argument, for normalized histograms. (GH2710).
DataFrame.from_records now accepts not only dicts but any instance of the collections.Mapping ABC.
added option
display.mpl_style
providing a sleeker visual style for plots. Based on https://gist.github.com/huyng/816622 (GH3075).Treat boolean values as integers (values 1 and 0) for numeric operations. (GH2641)
to_html() now accepts an optional “escape” argument to control reserved HTML character escaping (enabled by default) and escapes
&
, in addition to<
and>
. (GH2919)
See the full release notes or issue tracker on GitHub for a complete list.
Contributors¶
A total of 50 people contributed patches to this release. People with a “+” by their names contributed a patch for the first time.
Adam Greenhall +
Alvaro Tejero-Cantero +
Andy Hayden
Brad Buran +
Chang She
Chapman Siu +
Chris Withers +
Christian Geier +
Christopher Whelan
Damien Garaud
Dan Birken
Dan Davison +
Dieter Vandenbussche
Dražen Lučanin +
Dražen Lučanin +
Garrett Drapala
Illia Polosukhin +
James Casbon +
Jeff Reback
Jeremy Wagner +
Jonathan Chambers +
K.-Michael Aye
Karmel Allison +
Loïc Estève +
Nicholaus E. Halecky +
Peter Prettenhofer +
Phillip Cloud +
Robert Gieseke +
Skipper Seabold
Spencer Lyon
Stephen Lin +
Thierry Moisan +
Thomas Kluyver
Tim Akinbo +
Vytautas Jancauskas
Vytautas Jančauskas +
Wes McKinney
Will Furnass +
Wouter Overmeire
anomrake +
davidjameshumphreys +
dengemann +
dieterv77 +
jreback
lexual +
stephenwlin +
thauck +
vytas +
waitingkuo +
y-p