What’s new in 0.23.0 (May 15, 2018)#
This is a major release from 0.22.0 and includes a number of API changes, deprecations, 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:
Check the API Changes and deprecations before updating.
Warning
Starting January 1, 2019, pandas feature releases will support Python 3 only. See Dropping Python 2.7 for more.
What’s new in v0.23.0
New features#
JSON read/write roundtrippable with orient='table'
#
A DataFrame
can now be written to and subsequently read back via JSON while preserving metadata through usage of the orient='table'
argument (see GH 18912 and GH 9146). Previously, none of the available orient
values guaranteed the preservation of dtypes and index names, amongst other metadata.
In [1]: df = pd.DataFrame({'foo': [1, 2, 3, 4],
...: 'bar': ['a', 'b', 'c', 'd'],
...: 'baz': pd.date_range('20180101', freq='d', periods=4),
...: 'qux': pd.Categorical(['a', 'b', 'c', 'c'])},
...: index=pd.Index(range(4), name='idx'))
...:
In [2]: df
Out[2]:
foo bar baz qux
idx
0 1 a 20180101 a
1 2 b 20180102 b
2 3 c 20180103 c
3 4 d 20180104 c
[4 rows x 4 columns]
In [3]: df.dtypes
Out[3]:
foo int64
bar object
baz datetime64[ns]
qux category
Length: 4, dtype: object
In [4]: df.to_json('test.json', orient='table')
In [5]: new_df = pd.read_json('test.json', orient='table')
In [6]: new_df
Out[6]:
foo bar baz qux
idx
0 1 a 20180101 a
1 2 b 20180102 b
2 3 c 20180103 c
3 4 d 20180104 c
[4 rows x 4 columns]
In [7]: new_df.dtypes
Out[7]:
foo int64
bar object
baz datetime64[ns]
qux category
Length: 4, dtype: object
Please note that the string index
is not supported with the round trip format, as it is used by default in write_json
to indicate a missing index name.
In [8]: df.index.name = 'index'
In [9]: df.to_json('test.json', orient='table')
In [10]: new_df = pd.read_json('test.json', orient='table')
In [11]: new_df
Out[11]:
foo bar baz qux
0 1 a 20180101 a
1 2 b 20180102 b
2 3 c 20180103 c
3 4 d 20180104 c
[4 rows x 4 columns]
In [12]: new_df.dtypes
Out[12]:
foo int64
bar object
baz datetime64[ns]
qux category
Length: 4, dtype: object
Method .assign()
accepts dependent arguments#
The DataFrame.assign()
now accepts dependent keyword arguments for python version later than 3.6 (see also PEP 468). Later keyword arguments may now refer to earlier ones if the argument is a callable. See the
documentation here (GH 14207)
In [13]: df = pd.DataFrame({'A': [1, 2, 3]})
In [14]: df
Out[14]:
A
0 1
1 2
2 3
[3 rows x 1 columns]
In [15]: df.assign(B=df.A, C=lambda x: x['A'] + x['B'])
Out[15]:
A B C
0 1 1 2
1 2 2 4
2 3 3 6
[3 rows x 3 columns]
Warning
This may subtly change the behavior of your code when you’re
using .assign()
to update an existing column. Previously, callables
referring to other variables being updated would get the “old” values
Previous behavior:
In [2]: df = pd.DataFrame({"A": [1, 2, 3]})
In [3]: df.assign(A=lambda df: df.A + 1, C=lambda df: df.A * 1)
Out[3]:
A C
0 2 1
1 3 2
2 4 3
New behavior:
In [16]: df.assign(A=df.A + 1, C=lambda df: df.A * 1)
Out[16]:
A C
0 2 2
1 3 3
2 4 4
[3 rows x 2 columns]
Merging on a combination of columns and index levels#
Strings passed to DataFrame.merge()
as the on
, left_on
, and right_on
parameters may now refer to either column names or index level names.
This enables merging DataFrame
instances on a combination of index levels
and columns without resetting indexes. See the Merge on columns and
levels documentation section.
(GH 14355)
In [17]: left_index = pd.Index(['K0', 'K0', 'K1', 'K2'], name='key1')
In [18]: left = pd.DataFrame({'A': ['A0', 'A1', 'A2', 'A3'],
....: 'B': ['B0', 'B1', 'B2', 'B3'],
....: 'key2': ['K0', 'K1', 'K0', 'K1']},
....: index=left_index)
....:
In [19]: right_index = pd.Index(['K0', 'K1', 'K2', 'K2'], name='key1')
In [20]: right = pd.DataFrame({'C': ['C0', 'C1', 'C2', 'C3'],
....: 'D': ['D0', 'D1', 'D2', 'D3'],
....: 'key2': ['K0', 'K0', 'K0', 'K1']},
....: index=right_index)
....:
In [21]: left.merge(right, on=['key1', 'key2'])
Out[21]:
A B key2 C D
key1
K0 A0 B0 K0 C0 D0
K1 A2 B2 K0 C1 D1
K2 A3 B3 K1 C3 D3
[3 rows x 5 columns]
Sorting by a combination of columns and index levels#
Strings passed to DataFrame.sort_values()
as the by
parameter may
now refer to either column names or index level names. This enables sorting
DataFrame
instances by a combination of index levels and columns without
resetting indexes. See the Sorting by Indexes and Values documentation section.
(GH 14353)
# Build MultiIndex
In [22]: idx = pd.MultiIndex.from_tuples([('a', 1), ('a', 2), ('a', 2),
....: ('b', 2), ('b', 1), ('b', 1)])
....:
In [23]: idx.names = ['first', 'second']
# Build DataFrame
In [24]: df_multi = pd.DataFrame({'A': np.arange(6, 0, 1)},
....: index=idx)
....:
In [25]: df_multi
Out[25]:
A
first second
a 1 6
2 5
2 4
b 2 3
1 2
1 1
[6 rows x 1 columns]
# Sort by 'second' (index) and 'A' (column)
In [26]: df_multi.sort_values(by=['second', 'A'])
Out[26]:
A
first second
b 1 1
1 2
a 1 6
b 2 3
a 2 4
2 5
[6 rows x 1 columns]
Extending pandas with custom types (experimental)#
pandas now supports storing arraylike objects that aren’t necessarily 1D NumPy arrays as columns in a DataFrame or values in a Series. This allows thirdparty libraries to implement extensions to NumPy’s types, similar to how pandas implemented categoricals, datetimes with timezones, periods, and intervals.
As a demonstration, we’ll use cyberpandas, which provides an IPArray
type
for storing ip addresses.
In [1]: from cyberpandas import IPArray
In [2]: values = IPArray([
...: 0,
...: 3232235777,
...: 42540766452641154071740215577757643572
...: ])
...:
...:
IPArray
isn’t a normal 1D NumPy array, but because it’s a pandas
ExtensionArray
, it can be stored properly inside pandas’ containers.
In [3]: ser = pd.Series(values)
In [4]: ser
Out[4]:
0 0.0.0.0
1 192.168.1.1
2 2001:db8:85a3::8a2e:370:7334
dtype: ip
Notice that the dtype is ip
. The missing value semantics of the underlying
array are respected:
In [5]: ser.isna()
Out[5]:
0 True
1 False
2 False
dtype: bool
For more, see the extension types documentation. If you build an extension array, publicize it on the ecosystem page.
New observed
keyword for excluding unobserved categories in GroupBy
#
Grouping by a categorical includes the unobserved categories in the output.
When grouping by multiple categorical columns, this means you get the cartesian product of all the
categories, including combinations where there are no observations, which can result in a large
number of groups. We have added a keyword observed
to control this behavior, it defaults to
observed=False
for backwardcompatibility. (GH 14942, GH 8138, GH 15217, GH 17594, GH 8669, GH 20583, GH 20902)
In [27]: cat1 = pd.Categorical(["a", "a", "b", "b"],
....: categories=["a", "b", "z"], ordered=True)
....:
In [28]: cat2 = pd.Categorical(["c", "d", "c", "d"],
....: categories=["c", "d", "y"], ordered=True)
....:
In [29]: df = pd.DataFrame({"A": cat1, "B": cat2, "values": [1, 2, 3, 4]})
In [30]: df['C'] = ['foo', 'bar'] * 2
In [31]: df
Out[31]:
A B values C
0 a c 1 foo
1 a d 2 bar
2 b c 3 foo
3 b d 4 bar
[4 rows x 4 columns]
To show all values, the previous behavior:
In [32]: df.groupby(['A', 'B', 'C'], observed=False).count()
Out[32]:
values
A B C
a c bar 0
foo 1
d bar 1
foo 0
y bar 0
... ...
z c foo 0
d bar 0
foo 0
y bar 0
foo 0
[18 rows x 1 columns]
To show only observed values:
In [33]: df.groupby(['A', 'B', 'C'], observed=True).count()
Out[33]:
values
A B C
a c foo 1
d bar 1
b c foo 1
d bar 1
[4 rows x 1 columns]
For pivoting operations, this behavior is already controlled by the dropna
keyword:
In [34]: cat1 = pd.Categorical(["a", "a", "b", "b"],
....: categories=["a", "b", "z"], ordered=True)
....:
In [35]: cat2 = pd.Categorical(["c", "d", "c", "d"],
....: categories=["c", "d", "y"], ordered=True)
....:
In [36]: df = pd.DataFrame({"A": cat1, "B": cat2, "values": [1, 2, 3, 4]})
In [37]: df
Out[37]:
A B values
0 a c 1
1 a d 2
2 b c 3
3 b d 4
[4 rows x 3 columns]
In [38]: pd.pivot_table(df, values='values', index=['A', 'B'],
....: dropna=True)
....:
Out[38]:
values
A B
a c 1.0
d 2.0
b c 3.0
d 4.0
[4 rows x 1 columns]
In [39]: pd.pivot_table(df, values='values', index=['A', 'B'],
....: dropna=False)
....:
Out[39]:
values
A B
a c 1.0
d 2.0
y NaN
b c 3.0
d 4.0
y NaN
z c NaN
d NaN
y NaN
[9 rows x 1 columns]
Rolling/Expanding.apply() accepts raw=False
to pass a Series
to the function#
Series.rolling().apply()
, DataFrame.rolling().apply()
,
Series.expanding().apply()
, and DataFrame.expanding().apply()
have gained a raw=None
parameter.
This is similar to DataFame.apply()
. This parameter, if True
allows one to send a np.ndarray
to the applied function. If False
a Series
will be passed. The
default is None
, which preserves backward compatibility, so this will default to True
, sending an np.ndarray
.
In a future version the default will be changed to False
, sending a Series
. (GH 5071, GH 20584)
In [40]: s = pd.Series(np.arange(5), np.arange(5) + 1)
In [41]: s
Out[41]:
1 0
2 1
3 2
4 3
5 4
Length: 5, dtype: int64
Pass a Series
:
In [42]: s.rolling(2, min_periods=1).apply(lambda x: x.iloc[1], raw=False)
Out[42]:
1 0.0
2 1.0
3 2.0
4 3.0
5 4.0
Length: 5, dtype: float64
Mimic the original behavior of passing a ndarray:
In [43]: s.rolling(2, min_periods=1).apply(lambda x: x[1], raw=True)
Out[43]:
1 0.0
2 1.0
3 2.0
4 3.0
5 4.0
Length: 5, dtype: float64
DataFrame.interpolate
has gained the limit_area
kwarg#
DataFrame.interpolate()
has gained a limit_area
parameter to allow further control of which NaN
s are replaced.
Use limit_area='inside'
to fill only NaNs surrounded by valid values or use limit_area='outside'
to fill only NaN
s
outside the existing valid values while preserving those inside. (GH 16284) See the full documentation here.
In [44]: ser = pd.Series([np.nan, np.nan, 5, np.nan, np.nan,
....: np.nan, 13, np.nan, np.nan])
....:
In [45]: ser
Out[45]:
0 NaN
1 NaN
2 5.0
3 NaN
4 NaN
5 NaN
6 13.0
7 NaN
8 NaN
Length: 9, dtype: float64
Fill one consecutive inside value in both directions
In [46]: ser.interpolate(limit_direction='both', limit_area='inside', limit=1)
Out[46]:
0 NaN
1 NaN
2 5.0
3 7.0
4 NaN
5 11.0
6 13.0
7 NaN
8 NaN
Length: 9, dtype: float64
Fill all consecutive outside values backward
In [47]: ser.interpolate(limit_direction='backward', limit_area='outside')
Out[47]:
0 5.0
1 5.0
2 5.0
3 NaN
4 NaN
5 NaN
6 13.0
7 NaN
8 NaN
Length: 9, dtype: float64
Fill all consecutive outside values in both directions
In [48]: ser.interpolate(limit_direction='both', limit_area='outside')
Out[48]:
0 5.0
1 5.0
2 5.0
3 NaN
4 NaN
5 NaN
6 13.0
7 13.0
8 13.0
Length: 9, dtype: float64
Function get_dummies
now supports dtype
argument#
The get_dummies()
now accepts a dtype
argument, which specifies a dtype for the new columns. The default remains uint8. (GH 18330)
In [49]: df = pd.DataFrame({'a': [1, 2], 'b': [3, 4], 'c': [5, 6]})
In [50]: pd.get_dummies(df, columns=['c']).dtypes
Out[50]:
a int64
b int64
c_5 bool
c_6 bool
Length: 4, dtype: object
In [51]: pd.get_dummies(df, columns=['c'], dtype=bool).dtypes
Out[51]:
a int64
b int64
c_5 bool
c_6 bool
Length: 4, dtype: object
Timedelta mod method#
mod
(%) and divmod
operations are now defined on Timedelta
objects
when operating with either timedeltalike or with numeric arguments.
See the documentation here. (GH 19365)
In [52]: td = pd.Timedelta(hours=37)
In [53]: td % pd.Timedelta(minutes=45)
Out[53]: Timedelta('0 days 00:15:00')
Method .rank()
handles inf
values when NaN
are present#
In previous versions, .rank()
would assign inf
elements NaN
as their ranks. Now ranks are calculated properly. (GH 6945)
In [54]: s = pd.Series([np.inf, 0, 1, np.nan, np.inf])
In [55]: s
Out[55]:
0 inf
1 0.0
2 1.0
3 NaN
4 inf
Length: 5, dtype: float64
Previous behavior:
In [11]: s.rank()
Out[11]:
0 1.0
1 2.0
2 3.0
3 NaN
4 NaN
dtype: float64
Current behavior:
In [56]: s.rank()
Out[56]:
0 1.0
1 2.0
2 3.0
3 NaN
4 4.0
Length: 5, dtype: float64
Furthermore, previously if you rank inf
or inf
values together with NaN
values, the calculation won’t distinguish NaN
from infinity when using ‘top’ or ‘bottom’ argument.
In [57]: s = pd.Series([np.nan, np.nan, np.inf, np.inf])
In [58]: s
Out[58]:
0 NaN
1 NaN
2 inf
3 inf
Length: 4, dtype: float64
Previous behavior:
In [15]: s.rank(na_option='top')
Out[15]:
0 2.5
1 2.5
2 2.5
3 2.5
dtype: float64
Current behavior:
In [59]: s.rank(na_option='top')
Out[59]:
0 1.5
1 1.5
2 3.5
3 3.5
Length: 4, dtype: float64
These bugs were squashed:
Bug in
DataFrame.rank()
andSeries.rank()
whenmethod='dense'
andpct=True
in which percentile ranks were not being used with the number of distinct observations (GH 15630)Bug in
Series.rank()
andDataFrame.rank()
whenascending='False'
failed to return correct ranks for infinity ifNaN
were present (GH 19538)Bug in
DataFrameGroupBy.rank()
where ranks were incorrect when both infinity andNaN
were present (GH 20561)
Series.str.cat
has gained the join
kwarg#
Previously, Series.str.cat()
did not – in contrast to most of pandas
– align Series
on their index before concatenation (see GH 18657).
The method has now gained a keyword join
to control the manner of alignment, see examples below and here.
In v.0.23 join
will default to None (meaning no alignment), but this default will change to 'left'
in a future version of pandas.
In [60]: s = pd.Series(['a', 'b', 'c', 'd'])
In [61]: t = pd.Series(['b', 'd', 'e', 'c'], index=[1, 3, 4, 2])
In [62]: s.str.cat(t)
Out[62]:
0 NaN
1 bb
2 cc
3 dd
Length: 4, dtype: object
In [63]: s.str.cat(t, join='left', na_rep='')
Out[63]:
0 a
1 bb
2 cc
3 dd
Length: 4, dtype: object
Furthermore, Series.str.cat()
now works for CategoricalIndex
as well (previously raised a ValueError
; see GH 20842).
DataFrame.astype
performs columnwise conversion to Categorical
#
DataFrame.astype()
can now perform columnwise conversion to Categorical
by supplying the string 'category'
or
a CategoricalDtype
. Previously, attempting this would raise a NotImplementedError
. See the
Object creation section of the documentation for more details and examples. (GH 12860, GH 18099)
Supplying the string 'category'
performs columnwise conversion, with only labels appearing in a given column set as categories:
In [64]: df = pd.DataFrame({'A': list('abca'), 'B': list('bccd')})
In [65]: df = df.astype('category')
In [66]: df['A'].dtype
Out[66]: CategoricalDtype(categories=['a', 'b', 'c'], ordered=False, categories_dtype=object)
In [67]: df['B'].dtype
Out[67]: CategoricalDtype(categories=['b', 'c', 'd'], ordered=False, categories_dtype=object)
Supplying a CategoricalDtype
will make the categories in each column consistent with the supplied dtype:
In [68]: from pandas.api.types import CategoricalDtype
In [69]: df = pd.DataFrame({'A': list('abca'), 'B': list('bccd')})
In [70]: cdt = CategoricalDtype(categories=list('abcd'), ordered=True)
In [71]: df = df.astype(cdt)
In [72]: df['A'].dtype
Out[72]: CategoricalDtype(categories=['a', 'b', 'c', 'd'], ordered=True, categories_dtype=object)
In [73]: df['B'].dtype
Out[73]: CategoricalDtype(categories=['a', 'b', 'c', 'd'], ordered=True, categories_dtype=object)
Other enhancements#
Unary
+
now permitted forSeries
andDataFrame
as numeric operator (GH 16073)Better support for
to_excel()
output with thexlsxwriter
engine. (GH 16149)pandas.tseries.frequencies.to_offset()
now accepts leading ‘+’ signs e.g. ‘+1h’. (GH 18171)MultiIndex.unique()
now supports thelevel=
argument, to get unique values from a specific index level (GH 17896)pandas.io.formats.style.Styler
now has methodhide_index()
to determine whether the index will be rendered in output (GH 14194)pandas.io.formats.style.Styler
now has methodhide_columns()
to determine whether columns will be hidden in output (GH 14194)Improved wording of
ValueError
raised into_datetime()
whenunit=
is passed with a nonconvertible value (GH 14350)Series.fillna()
now accepts a Series or a dict as avalue
for a categorical dtype (GH 17033)pandas.read_clipboard()
updated to use qtpy, falling back to PyQt5 and then PyQt4, adding compatibility with Python3 and multiple pythonqt bindings (GH 17722)Improved wording of
ValueError
raised inread_csv()
when theusecols
argument cannot match all columns. (GH 17301)DataFrame.corrwith()
now silently drops nonnumeric columns when passed a Series. Before, an exception was raised (GH 18570).IntervalIndex
now supports time zone awareInterval
objects (GH 18537, GH 18538)Series()
/DataFrame()
tab completion also returns identifiers in the first level of aMultiIndex()
. (GH 16326)read_excel()
has gained thenrows
parameter (GH 16645)DataFrame.append()
can now in more cases preserve the type of the calling dataframe’s columns (e.g. if both areCategoricalIndex
) (GH 18359)DataFrame.to_json()
andSeries.to_json()
now accept anindex
argument which allows the user to exclude the index from the JSON output (GH 17394)IntervalIndex.to_tuples()
has gained thena_tuple
parameter to control whether NA is returned as a tuple of NA, or NA itself (GH 18756)Categorical.rename_categories
,CategoricalIndex.rename_categories
andSeries.cat.rename_categories
can now take a callable as their argument (GH 18862)Interval
andIntervalIndex
have gained alength
attribute (GH 18789)Resampler
objects now have a functioningpipe
method. Previously, calls topipe
were diverted to themean
method (GH 17905).is_scalar()
now returnsTrue
forDateOffset
objects (GH 18943).DataFrame.pivot()
now accepts a list for thevalues=
kwarg (GH 17160).Added
pandas.api.extensions.register_dataframe_accessor()
,pandas.api.extensions.register_series_accessor()
, andpandas.api.extensions.register_index_accessor()
, accessor for libraries downstream of pandas to register custom accessors like.cat
on pandas objects. See Registering Custom Accessors for more (GH 14781).IntervalIndex.astype
now supports conversions between subtypes when passed anIntervalDtype
(GH 19197)IntervalIndex
and its associated constructor methods (from_arrays
,from_breaks
,from_tuples
) have gained adtype
parameter (GH 19262)Added
pandas.core.groupby.SeriesGroupBy.is_monotonic_increasing()
andpandas.core.groupby.SeriesGroupBy.is_monotonic_decreasing()
(GH 17015)For subclassed
DataFrames
,DataFrame.apply()
will now preserve theSeries
subclass (if defined) when passing the data to the applied function (GH 19822)DataFrame.from_dict()
now accepts acolumns
argument that can be used to specify the column names whenorient='index'
is used (GH 18529)Added option
display.html.use_mathjax
so MathJax can be disabled when rendering tables inJupyter
notebooks (GH 19856, GH 19824)DataFrame.replace()
now supports themethod
parameter, which can be used to specify the replacement method whento_replace
is a scalar, list or tuple andvalue
isNone
(GH 19632)Timestamp.month_name()
,DatetimeIndex.month_name()
, andSeries.dt.month_name()
are now available (GH 12805)Timestamp.day_name()
andDatetimeIndex.day_name()
are now available to return day names with a specified locale (GH 12806)DataFrame.to_sql()
now performs a multivalue insert if the underlying connection supports itk rather than inserting row by row.SQLAlchemy
dialects supporting multivalue inserts include:mysql
,postgresql
,sqlite
and any dialect withsupports_multivalues_insert
. (GH 14315, GH 8953)read_html()
now accepts adisplayed_only
keyword argument to controls whether or not hidden elements are parsed (True
by default) (GH 20027)read_html()
now reads all<tbody>
elements in a<table>
, not just the first. (GH 20690)quantile()
andquantile()
now accept theinterpolation
keyword,linear
by default (GH 20497)zip compression is supported via
compression=zip
inDataFrame.to_pickle()
,Series.to_pickle()
,DataFrame.to_csv()
,Series.to_csv()
,DataFrame.to_json()
,Series.to_json()
. (GH 17778)WeekOfMonth
constructor now supportsn=0
(GH 20517).DataFrame
andSeries
now support matrix multiplication (@
) operator (GH 10259) for Python>=3.5Updated
DataFrame.to_gbq()
andpandas.read_gbq()
signature and documentation to reflect changes from the pandasgbq library version 0.4.0. Adds intersphinx mapping to pandasgbq library. (GH 20564)Added new writer for exporting Stata dta files in version 117,
StataWriter117
. This format supports exporting strings with lengths up to 2,000,000 characters (GH 16450)to_hdf()
andread_hdf()
now accept anerrors
keyword argument to control encoding error handling (GH 20835)cut()
has gained theduplicates='raise''drop'
option to control whether to raise on duplicated edges (GH 20947)date_range()
,timedelta_range()
, andinterval_range()
now return a linearly spaced index ifstart
,stop
, andperiods
are specified, butfreq
is not. (GH 20808, GH 20983, GH 20976)
Backwards incompatible API changes#
Dependencies have increased minimum versions#
We have updated our minimum supported versions of dependencies (GH 15184). If installed, we now require:
Package 
Minimum Version 
Required 
Issue 

pythondateutil 
2.5.0 
X 

openpyxl 
2.4.0 

beautifulsoup4 
4.2.1 

setuptools 
24.2.0 
Instantiation from dicts preserves dict insertion order for Python 3.6+#
Until Python 3.6, dicts in Python had no formally defined ordering. For Python
version 3.6 and later, dicts are ordered by insertion order, see
PEP 468.
pandas will use the dict’s insertion order, when creating a Series
or
DataFrame
from a dict and you’re using Python version 3.6 or
higher. (GH 19884)
Previous behavior (and current behavior if on Python < 3.6):
In [16]: pd.Series({'Income': 2000,
....: 'Expenses': 1500,
....: 'Taxes': 200,
....: 'Net result': 300})
Out[16]:
Expenses 1500
Income 2000
Net result 300
Taxes 200
dtype: int64
Note the Series above is ordered alphabetically by the index values.
New behavior (for Python >= 3.6):
In [74]: pd.Series({'Income': 2000,
....: 'Expenses': 1500,
....: 'Taxes': 200,
....: 'Net result': 300})
....:
Out[74]:
Income 2000
Expenses 1500
Taxes 200
Net result 300
Length: 4, dtype: int64
Notice that the Series is now ordered by insertion order. This new behavior is
used for all relevant pandas types (Series
, DataFrame
, SparseSeries
and SparseDataFrame
).
If you wish to retain the old behavior while using Python >= 3.6, you can use
.sort_index()
:
In [75]: pd.Series({'Income': 2000,
....: 'Expenses': 1500,
....: 'Taxes': 200,
....: 'Net result': 300}).sort_index()
....:
Out[75]:
Expenses 1500
Income 2000
Net result 300
Taxes 200
Length: 4, dtype: int64
Deprecate Panel#
Panel
was deprecated in the 0.20.x release, showing as a DeprecationWarning
. Using Panel
will now show a FutureWarning
. The recommended way to represent 3D data are
with a MultiIndex
on a DataFrame
via the to_frame()
or with the xarray package. pandas
provides a to_xarray()
method to automate this conversion (GH 13563, GH 18324).
In [75]: import pandas._testing as tm
In [76]: p = tm.makePanel()
In [77]: p
Out[77]:
<class 'pandas.core.panel.Panel'>
Dimensions: 3 (items) x 3 (major_axis) x 4 (minor_axis)
Items axis: ItemA to ItemC
Major_axis axis: 20000103 00:00:00 to 20000105 00:00:00
Minor_axis axis: A to D
Convert to a MultiIndex DataFrame
In [78]: p.to_frame()
Out[78]:
ItemA ItemB ItemC
major minor
20000103 A 0.469112 0.721555 0.404705
B 1.135632 0.271860 1.039268
C 0.119209 0.276232 1.344312
D 2.104569 0.113648 0.109050
20000104 A 0.282863 0.706771 0.577046
B 1.212112 0.424972 0.370647
C 1.044236 1.087401 0.844885
D 0.494929 1.478427 1.643563
20000105 A 1.509059 1.039575 1.715002
B 0.173215 0.567020 1.157892
C 0.861849 0.673690 1.075770
D 1.071804 0.524988 1.469388
[12 rows x 3 columns]
Convert to an xarray DataArray
In [79]: p.to_xarray()
Out[79]:
<xarray.DataArray (items: 3, major_axis: 3, minor_axis: 4)>
array([[[ 0.469112, 1.135632, 0.119209, 2.104569],
[0.282863, 1.212112, 1.044236, 0.494929],
[1.509059, 0.173215, 0.861849, 1.071804]],
[[ 0.721555, 0.27186 , 0.276232, 0.113648],
[0.706771, 0.424972, 1.087401, 1.478427],
[1.039575, 0.56702 , 0.67369 , 0.524988]],
[[ 0.404705, 1.039268, 1.344312, 0.10905 ],
[ 0.577046, 0.370647, 0.844885, 1.643563],
[1.715002, 1.157892, 1.07577 , 1.469388]]])
Coordinates:
* items (items) object 'ItemA' 'ItemB' 'ItemC'
* major_axis (major_axis) datetime64[ns] 20000103 20000104 20000105
* minor_axis (minor_axis) object 'A' 'B' 'C' 'D'
pandas.core.common removals#
The following error & warning messages are removed from pandas.core.common
(GH 13634, GH 19769):
PerformanceWarning
UnsupportedFunctionCall
UnsortedIndexError
AbstractMethodError
These are available from import from pandas.errors
(since 0.19.0).
Changes to make output of DataFrame.apply
consistent#
DataFrame.apply()
was inconsistent when applying an arbitrary userdefinedfunction that returned a listlike with axis=1
. Several bugs and inconsistencies
are resolved. If the applied function returns a Series, then pandas will return a DataFrame; otherwise a Series will be returned, this includes the case
where a listlike (e.g. tuple
or list
is returned) (GH 16353, GH 17437, GH 17970, GH 17348, GH 17892, GH 18573,
GH 17602, GH 18775, GH 18901, GH 18919).
In [76]: df = pd.DataFrame(np.tile(np.arange(3), 6).reshape(6, 1) + 1,
....: columns=['A', 'B', 'C'])
....:
In [77]: df
Out[77]:
A B C
0 1 2 3
1 1 2 3
2 1 2 3
3 1 2 3
4 1 2 3
5 1 2 3
[6 rows x 3 columns]
Previous behavior: if the returned shape happened to match the length of original columns, this would return a DataFrame
.
If the return shape did not match, a Series
with lists was returned.
In [3]: df.apply(lambda x: [1, 2, 3], axis=1)
Out[3]:
A B C
0 1 2 3
1 1 2 3
2 1 2 3
3 1 2 3
4 1 2 3
5 1 2 3
In [4]: df.apply(lambda x: [1, 2], axis=1)
Out[4]:
0 [1, 2]
1 [1, 2]
2 [1, 2]
3 [1, 2]
4 [1, 2]
5 [1, 2]
dtype: object
New behavior: When the applied function returns a listlike, this will now always return a Series
.
In [78]: df.apply(lambda x: [1, 2, 3], axis=1)
Out[78]:
0 [1, 2, 3]
1 [1, 2, 3]
2 [1, 2, 3]
3 [1, 2, 3]
4 [1, 2, 3]
5 [1, 2, 3]
Length: 6, dtype: object
In [79]: df.apply(lambda x: [1, 2], axis=1)
Out[79]:
0 [1, 2]
1 [1, 2]
2 [1, 2]
3 [1, 2]
4 [1, 2]
5 [1, 2]
Length: 6, dtype: object
To have expanded columns, you can use result_type='expand'
In [80]: df.apply(lambda x: [1, 2, 3], axis=1, result_type='expand')
Out[80]:
0 1 2
0 1 2 3
1 1 2 3
2 1 2 3
3 1 2 3
4 1 2 3
5 1 2 3
[6 rows x 3 columns]
To broadcast the result across the original columns (the old behaviour for
listlikes of the correct length), you can use result_type='broadcast'
.
The shape must match the original columns.
In [81]: df.apply(lambda x: [1, 2, 3], axis=1, result_type='broadcast')
Out[81]:
A B C
0 1 2 3
1 1 2 3
2 1 2 3
3 1 2 3
4 1 2 3
5 1 2 3
[6 rows x 3 columns]
Returning a Series
allows one to control the exact return structure and column names:
In [82]: df.apply(lambda x: pd.Series([1, 2, 3], index=['D', 'E', 'F']), axis=1)
Out[82]:
D E F
0 1 2 3
1 1 2 3
2 1 2 3
3 1 2 3
4 1 2 3
5 1 2 3
[6 rows x 3 columns]
Concatenation will no longer sort#
In a future version of pandas pandas.concat()
will no longer sort the nonconcatenation axis when it is not already aligned.
The current behavior is the same as the previous (sorting), but now a warning is issued when sort
is not specified and the nonconcatenation axis is not aligned (GH 4588).
In [83]: df1 = pd.DataFrame({"a": [1, 2], "b": [1, 2]}, columns=['b', 'a'])
In [84]: df2 = pd.DataFrame({"a": [4, 5]})
In [85]: pd.concat([df1, df2])
Out[85]:
b a
0 1.0 1
1 2.0 2
0 NaN 4
1 NaN 5
[4 rows x 2 columns]
To keep the previous behavior (sorting) and silence the warning, pass sort=True
In [86]: pd.concat([df1, df2], sort=True)
Out[86]:
a b
0 1 1.0
1 2 2.0
0 4 NaN
1 5 NaN
[4 rows x 2 columns]
To accept the future behavior (no sorting), pass sort=False
Note that this change also applies to DataFrame.append()
, which has also received a sort
keyword for controlling this behavior.
Build changes#
Index division by zero fills correctly#
Division operations on Index
and subclasses will now fill division of positive numbers by zero with np.inf
, division of negative numbers by zero with np.inf
and 0 / 0
with np.nan
. This matches existing Series
behavior. (GH 19322, GH 19347)
Previous behavior:
In [6]: index = pd.Int64Index([1, 0, 1])
In [7]: index / 0
Out[7]: Int64Index([0, 0, 0], dtype='int64')
# Previous behavior yielded different results depending on the type of zero in the divisor
In [8]: index / 0.0
Out[8]: Float64Index([inf, nan, inf], dtype='float64')
In [9]: index = pd.UInt64Index([0, 1])
In [10]: index / np.array([0, 0], dtype=np.uint64)
Out[10]: UInt64Index([0, 0], dtype='uint64')
In [11]: pd.RangeIndex(1, 5) / 0
ZeroDivisionError: integer division or modulo by zero
Current behavior:
In [12]: index = pd.Int64Index([1, 0, 1])
# division by zero gives infinity where negative,
# +infinity where positive, and NaN for 0 / 0
In [13]: index / 0
# The result of division by zero should not depend on
# whether the zero is int or float
In [14]: index / 0.0
In [15]: index = pd.UInt64Index([0, 1])
In [16]: index / np.array([0, 0], dtype=np.uint64)
In [17]: pd.RangeIndex(1, 5) / 0
Extraction of matching patterns from strings#
By default, extracting matching patterns from strings with str.extract()
used to return a
Series
if a single group was being extracted (a DataFrame
if more than one group was
extracted). As of pandas 0.23.0 str.extract()
always returns a DataFrame
, unless
expand
is set to False
. Finally, None
was an accepted value for
the expand
parameter (which was equivalent to False
), but now raises a ValueError
. (GH 11386)
Previous behavior:
In [1]: s = pd.Series(['number 10', '12 eggs'])
In [2]: extracted = s.str.extract(r'.*(\d\d).*')
In [3]: extracted
Out [3]:
0 10
1 12
dtype: object
In [4]: type(extracted)
Out [4]:
pandas.core.series.Series
New behavior:
In [87]: s = pd.Series(['number 10', '12 eggs'])
In [88]: extracted = s.str.extract(r'.*(\d\d).*')
In [89]: extracted
Out[89]:
0
0 10
1 12
[2 rows x 1 columns]
In [90]: type(extracted)
Out[90]: pandas.core.frame.DataFrame
To restore previous behavior, simply set expand
to False
:
In [91]: s = pd.Series(['number 10', '12 eggs'])
In [92]: extracted = s.str.extract(r'.*(\d\d).*', expand=False)
In [93]: extracted
Out[93]:
0 10
1 12
Length: 2, dtype: object
In [94]: type(extracted)
Out[94]: pandas.core.series.Series
Default value for the ordered
parameter of CategoricalDtype
#
The default value of the ordered
parameter for CategoricalDtype
has changed from False
to None
to allow updating of categories
without impacting ordered
. Behavior should remain consistent for downstream objects, such as Categorical
(GH 18790)
In previous versions, the default value for the ordered
parameter was False
. This could potentially lead to the ordered
parameter unintentionally being changed from True
to False
when users attempt to update categories
if ordered
is not explicitly specified, as it would silently default to False
. The new behavior for ordered=None
is to retain the existing value of ordered
.
New behavior:
In [2]: from pandas.api.types import CategoricalDtype
In [3]: cat = pd.Categorical(list('abcaba'), ordered=True, categories=list('cba'))
In [4]: cat
Out[4]:
[a, b, c, a, b, a]
Categories (3, object): [c < b < a]
In [5]: cdt = CategoricalDtype(categories=list('cbad'))
In [6]: cat.astype(cdt)
Out[6]:
[a, b, c, a, b, a]
Categories (4, object): [c < b < a < d]
Notice in the example above that the converted Categorical
has retained ordered=True
. Had the default value for ordered
remained as False
, the converted Categorical
would have become unordered, despite ordered=False
never being explicitly specified. To change the value of ordered
, explicitly pass it to the new dtype, e.g. CategoricalDtype(categories=list('cbad'), ordered=False)
.
Note that the unintentional conversion of ordered
discussed above did not arise in previous versions due to separate bugs that prevented astype
from doing any type of category to category conversion (GH 10696, GH 18593). These bugs have been fixed in this release, and motivated changing the default value of ordered
.
Better prettyprinting of DataFrames in a terminal#
Previously, the default value for the maximum number of columns was
pd.options.display.max_columns=20
. This meant that relatively wide data
frames would not fit within the terminal width, and pandas would introduce line
breaks to display these 20 columns. This resulted in an output that was
relatively difficult to read:
If Python runs in a terminal, the maximum number of columns is now determined
automatically so that the printed data frame fits within the current terminal
width (pd.options.display.max_columns=0
) (GH 17023). If Python runs
as a Jupyter kernel (such as the Jupyter QtConsole or a Jupyter notebook, as
well as in many IDEs), this value cannot be inferred automatically and is thus
set to 20
as in previous versions. In a terminal, this results in a much
nicer output:
Note that if you don’t like the new default, you can always set this option yourself. To revert to the old setting, you can run this line:
pd.options.display.max_columns = 20
Datetimelike API changes#
The default
Timedelta
constructor now accepts anISO 8601 Duration
string as an argument (GH 19040)Subtracting
NaT
from aSeries
withdtype='datetime64[ns]'
returns aSeries
withdtype='timedelta64[ns]'
instead ofdtype='datetime64[ns]'
(GH 18808)Addition or subtraction of
NaT
fromTimedeltaIndex
will returnTimedeltaIndex
instead ofDatetimeIndex
(GH 19124)DatetimeIndex.shift()
andTimedeltaIndex.shift()
will now raiseNullFrequencyError
(which subclassesValueError
, which was raised in older versions) when the index object frequency isNone
(GH 19147)Addition and subtraction of
NaN
from aSeries
withdtype='timedelta64[ns]'
will raise aTypeError
instead of treating theNaN
asNaT
(GH 19274)NaT
division withdatetime.timedelta
will now returnNaN
instead of raising (GH 17876)Operations between a
Series
with dtypedtype='datetime64[ns]'
and aPeriodIndex
will correctly raisesTypeError
(GH 18850)Subtraction of
Series
with timezoneawaredtype='datetime64[ns]'
with mismatched timezones will raiseTypeError
instead ofValueError
(GH 18817)Timestamp
will no longer silently ignore unused or invalidtz
ortzinfo
keyword arguments (GH 17690)Timestamp
will no longer silently ignore invalidfreq
arguments (GH 5168)CacheableOffset
andWeekDay
are no longer available in thepandas.tseries.offsets
module (GH 17830)pandas.tseries.frequencies.get_freq_group()
andpandas.tseries.frequencies.DAYS
are removed from the public API (GH 18034)Series.truncate()
andDataFrame.truncate()
will raise aValueError
if the index is not sorted instead of an unhelpfulKeyError
(GH 17935)Series.first
andDataFrame.first
will now raise aTypeError
rather thanNotImplementedError
when index is not aDatetimeIndex
(GH 20725).Series.last
andDataFrame.last
will now raise aTypeError
rather thanNotImplementedError
when index is not aDatetimeIndex
(GH 20725).Restricted
DateOffset
keyword arguments. Previously,DateOffset
subclasses allowed arbitrary keyword arguments which could lead to unexpected behavior. Now, only valid arguments will be accepted. (GH 17176, GH 18226).pandas.merge()
provides a more informative error message when trying to merge on timezoneaware and timezonenaive columns (GH 15800)For
DatetimeIndex
andTimedeltaIndex
withfreq=None
, addition or subtraction of integerdtyped array orIndex
will raiseNullFrequencyError
instead ofTypeError
(GH 19895)Timestamp
constructor now accepts ananosecond
keyword or positional argument (GH 18898)DatetimeIndex
will now raise anAttributeError
when thetz
attribute is set after instantiation (GH 3746)DatetimeIndex
with apytz
timezone will now return a consistentpytz
timezone (GH 18595)
Other API changes#
Series.astype()
andIndex.astype()
with an incompatible dtype will now raise aTypeError
rather than aValueError
(GH 18231)Series
construction with anobject
dtyped tzaware datetime anddtype=object
specified, will now return anobject
dtypedSeries
, previously this would infer the datetime dtype (GH 18231)A
Series
ofdtype=category
constructed from an emptydict
will now have categories ofdtype=object
rather thandtype=float64
, consistently with the case in which an empty list is passed (GH 18515)AllNaN levels in a
MultiIndex
are now assignedfloat
rather thanobject
dtype, promoting consistency withIndex
(GH 17929).Levels names of a
MultiIndex
(when not None) are now required to be unique: trying to create aMultiIndex
with repeated names will raise aValueError
(GH 18872)Both construction and renaming of
Index
/MultiIndex
with nonhashablename
/names
will now raiseTypeError
(GH 20527)Index.map()
can now acceptSeries
and dictionary input objects (GH 12756, GH 18482, GH 18509).DataFrame.unstack()
will now default to filling withnp.nan
forobject
columns. (GH 12815)IntervalIndex
constructor will raise if theclosed
parameter conflicts with how the input data is inferred to be closed (GH 18421)Inserting missing values into indexes will work for all types of indexes and automatically insert the correct type of missing value (
NaN
,NaT
, etc.) regardless of the type passed in (GH 18295)When created with duplicate labels,
MultiIndex
now raises aValueError
. (GH 17464)Series.fillna()
now raises aTypeError
instead of aValueError
when passed a list, tuple or DataFrame as avalue
(GH 18293)pandas.DataFrame.merge()
no longer casts afloat
column toobject
when merging onint
andfloat
columns (GH 16572)pandas.merge()
now raises aValueError
when trying to merge on incompatible data types (GH 9780)The default NA value for
UInt64Index
has changed from 0 toNaN
, which impacts methods that mask with NA, such asUInt64Index.where()
(GH 18398)Refactored
setup.py
to usefind_packages
instead of explicitly listing out all subpackages (GH 18535)Rearranged the order of keyword arguments in
read_excel()
to align withread_csv()
(GH 16672)wide_to_long()
previously kept numericlike suffixes asobject
dtype. Now they are cast to numeric if possible (GH 17627)In
read_excel()
, thecomment
argument is now exposed as a named parameter (GH 18735)Rearranged the order of keyword arguments in
read_excel()
to align withread_csv()
(GH 16672)The options
html.border
andmode.use_inf_as_null
were deprecated in prior versions, these will now showFutureWarning
rather than aDeprecationWarning
(GH 19003)IntervalIndex
andIntervalDtype
no longer support categorical, object, and string subtypes (GH 19016)IntervalDtype
now returnsTrue
when compared against'interval'
regardless of subtype, andIntervalDtype.name
now returns'interval'
regardless of subtype (GH 18980)KeyError
now raises instead ofValueError
indrop()
,drop()
,drop()
,drop()
when dropping a nonexistent element in an axis with duplicates (GH 19186)Series.to_csv()
now accepts acompression
argument that works in the same way as thecompression
argument inDataFrame.to_csv()
(GH 18958)Set operations (union, difference…) on
IntervalIndex
with incompatible index types will now raise aTypeError
rather than aValueError
(GH 19329)DateOffset
objects render more simply, e.g.<DateOffset: days=1>
instead of<DateOffset: kwds={'days': 1}>
(GH 19403)Categorical.fillna
now validates itsvalue
andmethod
keyword arguments. It now raises when both or none are specified, matching the behavior ofSeries.fillna()
(GH 19682)pd.to_datetime('today')
now returns a datetime, consistent withpd.Timestamp('today')
; previouslypd.to_datetime('today')
returned a.normalized()
datetime (GH 19935)Series.str.replace()
now takes an optionalregex
keyword which, when set toFalse
, uses literal string replacement rather than regex replacement (GH 16808)DatetimeIndex.strftime()
andPeriodIndex.strftime()
now return anIndex
instead of a numpy array to be consistent with similar accessors (GH 20127)Constructing a Series from a list of length 1 no longer broadcasts this list when a longer index is specified (GH 19714, GH 20391).
DataFrame.to_dict()
withorient='index'
no longer casts int columns to float for a DataFrame with only int and float columns (GH 18580)A userdefinedfunction that is passed to
Series.rolling().aggregate()
,DataFrame.rolling().aggregate()
, or its expanding cousins, will now always be passed aSeries
, rather than anp.array
;.apply()
only has theraw
keyword, see here. This is consistent with the signatures of.aggregate()
across pandas (GH 20584)Rolling and Expanding types raise
NotImplementedError
upon iteration (GH 11704).
Deprecations#
Series.from_array
andSparseSeries.from_array
are deprecated. Use the normal constructorSeries(..)
andSparseSeries(..)
instead (GH 18213).DataFrame.as_matrix
is deprecated. UseDataFrame.values
instead (GH 18458).Series.asobject
,DatetimeIndex.asobject
,PeriodIndex.asobject
andTimeDeltaIndex.asobject
have been deprecated. Use.astype(object)
instead (GH 18572)Grouping by a tuple of keys now emits a
FutureWarning
and is deprecated. In the future, a tuple passed to'by'
will always refer to a single key that is the actual tuple, instead of treating the tuple as multiple keys. To retain the previous behavior, use a list instead of a tuple (GH 18314)Series.valid
is deprecated. UseSeries.dropna()
instead (GH 18800).read_excel()
has deprecated theskip_footer
parameter. Useskipfooter
instead (GH 18836)ExcelFile.parse()
has deprecatedsheetname
in favor ofsheet_name
for consistency withread_excel()
(GH 20920).The
is_copy
attribute is deprecated and will be removed in a future version (GH 18801).IntervalIndex.from_intervals
is deprecated in favor of theIntervalIndex
constructor (GH 19263)DataFrame.from_items
is deprecated. UseDataFrame.from_dict()
instead, orDataFrame.from_dict(OrderedDict())
if you wish to preserve the key order (GH 17320, GH 17312)Indexing a
MultiIndex
or aFloatIndex
with a list containing some missing keys will now show aFutureWarning
, which is consistent with other types of indexes (GH 17758).The
broadcast
parameter of.apply()
is deprecated in favor ofresult_type='broadcast'
(GH 18577)The
reduce
parameter of.apply()
is deprecated in favor ofresult_type='reduce'
(GH 18577)The
order
parameter offactorize()
is deprecated and will be removed in a future release (GH 19727)Timestamp.weekday_name
,DatetimeIndex.weekday_name
, andSeries.dt.weekday_name
are deprecated in favor ofTimestamp.day_name()
,DatetimeIndex.day_name()
, andSeries.dt.day_name()
(GH 12806)pandas.tseries.plotting.tsplot
is deprecated. UseSeries.plot()
instead (GH 18627)Index.summary()
is deprecated and will be removed in a future version (GH 18217)NDFrame.get_ftype_counts()
is deprecated and will be removed in a future version (GH 18243)The
convert_datetime64
parameter inDataFrame.to_records()
has been deprecated and will be removed in a future version. The NumPy bug motivating this parameter has been resolved. The default value for this parameter has also changed fromTrue
toNone
(GH 18160).Series.rolling().apply()
,DataFrame.rolling().apply()
,Series.expanding().apply()
, andDataFrame.expanding().apply()
have deprecated passing annp.array
by default. One will need to pass the newraw
parameter to be explicit about what is passed (GH 20584)The
data
,base
,strides
,flags
anditemsize
properties of theSeries
andIndex
classes have been deprecated and will be removed in a future version (GH 20419).DatetimeIndex.offset
is deprecated. UseDatetimeIndex.freq
instead (GH 20716)Floor division between an integer ndarray and a
Timedelta
is deprecated. Divide byTimedelta.value
instead (GH 19761)Setting
PeriodIndex.freq
(which was not guaranteed to work correctly) is deprecated. UsePeriodIndex.asfreq()
instead (GH 20678)Index.get_duplicates()
is deprecated and will be removed in a future version (GH 20239)The previous default behavior of negative indices in
Categorical.take
is deprecated. In a future version it will change from meaning missing values to meaning positional indices from the right. The future behavior is consistent withSeries.take()
(GH 20664).Passing multiple axes to the
axis
parameter inDataFrame.dropna()
has been deprecated and will be removed in a future version (GH 20987)
Removal of prior version deprecations/changes#
Warnings against the obsolete usage
Categorical(codes, categories)
, which were emitted for instance when the first two arguments toCategorical()
had different dtypes, and recommended the use ofCategorical.from_codes
, have now been removed (GH 8074)The
levels
andlabels
attributes of aMultiIndex
can no longer be set directly (GH 4039).pd.tseries.util.pivot_annual
has been removed (deprecated since v0.19). Usepivot_table
instead (GH 18370)pd.tseries.util.isleapyear
has been removed (deprecated since v0.19). Use.is_leap_year
property in Datetimelikes instead (GH 18370)pd.ordered_merge
has been removed (deprecated since v0.19). Usepd.merge_ordered
instead (GH 18459)The
SparseList
class has been removed (GH 14007)The
pandas.io.wb
andpandas.io.data
stub modules have been removed (GH 13735)Categorical.from_array
has been removed (GH 13854)The
freq
andhow
parameters have been removed from therolling
/expanding
/ewm
methods of DataFrame and Series (deprecated since v0.18). Instead, resample before calling the methods. (GH 18601 & GH 18668)DatetimeIndex.to_datetime
,Timestamp.to_datetime
,PeriodIndex.to_datetime
, andIndex.to_datetime
have been removed (GH 8254, GH 14096, GH 14113)read_csv()
has dropped theskip_footer
parameter (GH 13386)read_csv()
has dropped theas_recarray
parameter (GH 13373)read_csv()
has dropped thebuffer_lines
parameter (GH 13360)read_csv()
has dropped thecompact_ints
anduse_unsigned
parameters (GH 13323)The
Timestamp
class has dropped theoffset
attribute in favor offreq
(GH 13593)The
Series
,Categorical
, andIndex
classes have dropped thereshape
method (GH 13012)pandas.tseries.frequencies.get_standard_freq
has been removed in favor ofpandas.tseries.frequencies.to_offset(freq).rule_code
(GH 13874)The
freqstr
keyword has been removed frompandas.tseries.frequencies.to_offset
in favor offreq
(GH 13874)The
Panel4D
andPanelND
classes have been removed (GH 13776)The
Panel
class has dropped theto_long
andtoLong
methods (GH 19077)The options
display.line_with
anddisplay.height
are removed in favor ofdisplay.width
anddisplay.max_rows
respectively (GH 4391, GH 19107)The
labels
attribute of theCategorical
class has been removed in favor ofCategorical.codes
(GH 7768)The
flavor
parameter have been removed fromto_sql()
method (GH 13611)The modules
pandas.tools.hashing
andpandas.util.hashing
have been removed (GH 16223)The toplevel functions
pd.rolling_*
,pd.expanding_*
andpd.ewm*
have been removed (Deprecated since v0.18). Instead, use the DataFrame/Series methodsrolling
,expanding
andewm
(GH 18723)Imports from
pandas.core.common
for functions such asis_datetime64_dtype
are now removed. These are located inpandas.api.types
. (GH 13634, GH 19769)The
infer_dst
keyword inSeries.tz_localize()
,DatetimeIndex.tz_localize()
andDatetimeIndex
have been removed.infer_dst=True
is equivalent toambiguous='infer'
, andinfer_dst=False
toambiguous='raise'
(GH 7963).When
.resample()
was changed from an eager to a lazy operation, like.groupby()
in v0.18.0, we put in place compatibility (with aFutureWarning
), so operations would continue to work. This is now fully removed, so aResampler
will no longer forward compat operations (GH 20554)Remove long deprecated
axis=None
parameter from.replace()
(GH 20271)
Performance improvements#
Indexers on
Series
orDataFrame
no longer create a reference cycle (GH 17956)Added a keyword argument,
cache
, toto_datetime()
that improved the performance of converting duplicate datetime arguments (GH 11665)DateOffset
arithmetic performance is improved (GH 18218)Converting a
Series
ofTimedelta
objects to days, seconds, etc… sped up through vectorization of underlying methods (GH 18092)Improved performance of
.map()
with aSeries/dict
input (GH 15081)The overridden
Timedelta
properties of days, seconds and microseconds have been removed, leveraging their builtin Python versions instead (GH 18242)Series
construction will reduce the number of copies made of the input data in certain cases (GH 17449)Improved performance of
Series.dt.date()
andDatetimeIndex.date()
(GH 18058)Improved performance of
Series.dt.time()
andDatetimeIndex.time()
(GH 18461)Improved performance of
IntervalIndex.symmetric_difference()
(GH 18475)Improved performance of
DatetimeIndex
andSeries
arithmetic operations with BusinessMonth and BusinessQuarter frequencies (GH 18489)Series()
/DataFrame()
tab completion limits to 100 values, for better performance. (GH 18587)Improved performance of
DataFrame.median()
withaxis=1
when bottleneck is not installed (GH 16468)Improved performance of
MultiIndex.get_loc()
for large indexes, at the cost of a reduction in performance for small ones (GH 18519)Improved performance of
MultiIndex.remove_unused_levels()
when there are no unused levels, at the cost of a reduction in performance when there are (GH 19289)Improved performance of
Index.get_loc()
for nonunique indexes (GH 19478)Improved performance of pairwise
.rolling()
and.expanding()
with.cov()
and.corr()
operations (GH 17917)Improved performance of
pandas.core.groupby.GroupBy.rank()
(GH 15779)Improved performance of variable
.rolling()
on.min()
and.max()
(GH 19521)Improved performance of
pandas.core.groupby.GroupBy.ffill()
andpandas.core.groupby.GroupBy.bfill()
(GH 11296)Improved performance of
pandas.core.groupby.GroupBy.any()
andpandas.core.groupby.GroupBy.all()
(GH 15435)Improved performance of
pandas.core.groupby.GroupBy.pct_change()
(GH 19165)Improved performance of
Series.isin()
in the case of categorical dtypes (GH 20003)Improved performance of
getattr(Series, attr)
when the Series has certain index types. This manifested in slow printing of large Series with aDatetimeIndex
(GH 19764)Fixed a performance regression for
GroupBy.nth()
andGroupBy.last()
with some object columns (GH 19283)Improved performance of
pandas.core.arrays.Categorical.from_codes()
(GH 18501)
Documentation changes#
Thanks to all of the contributors who participated in the pandas Documentation Sprint, which took place on March 10th. We had about 500 participants from over 30 locations across the world. You should notice that many of the API docstrings have greatly improved.
There were too many simultaneous contributions to include a release note for each improvement, but this GitHub search should give you an idea of how many docstrings were improved.
Special thanks to Marc Garcia for organizing the sprint. For more information, read the NumFOCUS blogpost recapping the sprint.
Changed spelling of “numpy” to “NumPy”, and “python” to “Python”. (GH 19017)
Consistency when introducing code samples, using either colon or period. Rewrote some sentences for greater clarity, added more dynamic references to functions, methods and classes. (GH 18941, GH 18948, GH 18973, GH 19017)
Added a reference to
DataFrame.assign()
in the concatenate section of the merging documentation (GH 18665)
Bug fixes#
Categorical#
Warning
A class of bugs were introduced in pandas 0.21 with CategoricalDtype
that
affects the correctness of operations like merge
, concat
, and
indexing when comparing multiple unordered Categorical
arrays that have
the same categories, but in a different order. We highly recommend upgrading
or manually aligning your categories before doing these operations.
Bug in
Categorical.equals
returning the wrong result when comparing two unorderedCategorical
arrays with the same categories, but in a different order (GH 16603)Bug in
pandas.api.types.union_categoricals()
returning the wrong result when for unordered categoricals with the categories in a different order. This affectedpandas.concat()
with Categorical data (GH 19096).Bug in
pandas.merge()
returning the wrong result when joining on an unorderedCategorical
that had the same categories but in a different order (GH 19551)Bug in
CategoricalIndex.get_indexer()
returning the wrong result whentarget
was an unorderedCategorical
that had the same categories asself
but in a different order (GH 19551)Bug in
Index.astype()
with a categorical dtype where the resultant index is not converted to aCategoricalIndex
for all types of index (GH 18630)Bug in
Series.astype()
andCategorical.astype()
where an existing categorical data does not get updated (GH 10696, GH 18593)Bug in
Series.str.split()
withexpand=True
incorrectly raising an IndexError on empty strings (GH 20002).Bug in
Index
constructor withdtype=CategoricalDtype(...)
wherecategories
andordered
are not maintained (GH 19032)Bug in
Series
constructor with scalar anddtype=CategoricalDtype(...)
wherecategories
andordered
are not maintained (GH 19565)Bug in
Categorical.__iter__
not converting to Python types (GH 19909)Bug in
pandas.factorize()
returning the unique codes for theuniques
. This now returns aCategorical
with the same dtype as the input (GH 19721)Bug in
pandas.factorize()
including an item for missing values in theuniques
return value (GH 19721)Bug in
Series.take()
with categorical data interpreting1
inindices
as missing value markers, rather than the last element of the Series (GH 20664)
Datetimelike#
Bug in
Series.__sub__()
subtracting a nonnanosecondnp.datetime64
object from aSeries
gave incorrect results (GH 7996)Bug in
DatetimeIndex
,TimedeltaIndex
addition and subtraction of zerodimensional integer arrays gave incorrect results (GH 19012)Bug in
DatetimeIndex
andTimedeltaIndex
where adding or subtracting an arraylike ofDateOffset
objects either raised (np.array
,pd.Index
) or broadcast incorrectly (pd.Series
) (GH 18849)Bug in
Series.__add__()
adding Series with dtypetimedelta64[ns]
to a timezoneawareDatetimeIndex
incorrectly dropped timezone information (GH 13905)Adding a
Period
object to adatetime
orTimestamp
object will now correctly raise aTypeError
(GH 17983)Bug in
Timestamp
where comparison with an array ofTimestamp
objects would result in aRecursionError
(GH 15183)Bug in
Series
floordivision where operating on a scalartimedelta
raises an exception (GH 18846)Bug in
DatetimeIndex
where the repr was not showing highprecision time values at the end of a day (e.g., 23:59:59.999999999) (GH 19030)Bug in
.astype()
to nonns timedelta units would hold the incorrect dtype (GH 19176, GH 19223, GH 12425)Bug in subtracting
Series
fromNaT
incorrectly returningNaT
(GH 19158)Bug in
Series.truncate()
which raisesTypeError
with a monotonicPeriodIndex
(GH 17717)Bug in
pct_change()
usingperiods
andfreq
returned different length outputs (GH 7292)Bug in comparison of
DatetimeIndex
againstNone
ordatetime.date
objects raisingTypeError
for==
and!=
comparisons instead of allFalse
and allTrue
, respectively (GH 19301)Bug in
Timestamp
andto_datetime()
where a string representing a barely outofbounds timestamp would be incorrectly rounded down instead of raisingOutOfBoundsDatetime
(GH 19382)Bug in
Timestamp.floor()
DatetimeIndex.floor()
where time stamps far in the future and past were not rounded correctly (GH 19206)Bug in
to_datetime()
where passing an outofbounds datetime witherrors='coerce'
andutc=True
would raiseOutOfBoundsDatetime
instead of parsing toNaT
(GH 19612)Bug in
DatetimeIndex
andTimedeltaIndex
addition and subtraction where name of the returned object was not always set consistently. (GH 19744)Bug in
DatetimeIndex
andTimedeltaIndex
addition and subtraction where operations with numpy arrays raisedTypeError
(GH 19847)Bug in
DatetimeIndex
andTimedeltaIndex
where setting thefreq
attribute was not fully supported (GH 20678)
Timedelta#
Bug in
Timedelta.__mul__()
where multiplying byNaT
returnedNaT
instead of raising aTypeError
(GH 19819)Bug in
Series
withdtype='timedelta64[ns]'
where addition or subtraction ofTimedeltaIndex
had results cast todtype='int64'
(GH 17250)Bug in
Series
withdtype='timedelta64[ns]'
where addition or subtraction ofTimedeltaIndex
could return aSeries
with an incorrect name (GH 19043)Bug in
Timedelta.__floordiv__()
andTimedelta.__rfloordiv__()
dividing by many incompatible numpy objects was incorrectly allowed (GH 18846)Bug where dividing a scalar timedeltalike object with
TimedeltaIndex
performed the reciprocal operation (GH 19125)Bug in
TimedeltaIndex
where division by aSeries
would return aTimedeltaIndex
instead of aSeries
(GH 19042)Bug in
Timedelta.__add__()
,Timedelta.__sub__()
where adding or subtracting anp.timedelta64
object would return anothernp.timedelta64
instead of aTimedelta
(GH 19738)Bug in
Timedelta.__floordiv__()
,Timedelta.__rfloordiv__()
where operating with aTick
object would raise aTypeError
instead of returning a numeric value (GH 19738)Bug in
Period.asfreq()
where periods neardatetime(1, 1, 1)
could be converted incorrectly (GH 19643, GH 19834)Bug in
Timedelta.total_seconds()
causing precision errors, for exampleTimedelta('30S').total_seconds()==30.000000000000004
(GH 19458)Bug in
Timedelta.__rmod__()
where operating with anumpy.timedelta64
returned atimedelta64
object instead of aTimedelta
(GH 19820)Multiplication of
TimedeltaIndex
byTimedeltaIndex
will now raiseTypeError
instead of raisingValueError
in cases of length mismatch (GH 19333)Bug in indexing a
TimedeltaIndex
with anp.timedelta64
object which was raising aTypeError
(GH 20393)
Timezones#
Bug in creating a
Series
from an array that contains both tznaive and tzaware values will result in aSeries
whose dtype is tzaware instead of object (GH 16406)Bug in comparison of timezoneaware
DatetimeIndex
againstNaT
incorrectly raisingTypeError
(GH 19276)Bug in
DatetimeIndex.astype()
when converting between timezone aware dtypes, and converting from timezone aware to naive (GH 18951)Bug in comparing
DatetimeIndex
, which failed to raiseTypeError
when attempting to compare timezoneaware and timezonenaive datetimelike objects (GH 18162)Bug in localization of a naive, datetime string in a
Series
constructor with adatetime64[ns, tz]
dtype (GH 174151)Timestamp.replace()
will now handle Daylight Savings transitions gracefully (GH 18319)Bug in tzaware
DatetimeIndex
where addition/subtraction with aTimedeltaIndex
or array withdtype='timedelta64[ns]'
was incorrect (GH 17558)Bug in
DatetimeIndex.insert()
where insertingNaT
into a timezoneaware index incorrectly raised (GH 16357)Bug in
DataFrame
constructor, where tzaware Datetimeindex and a given column name will result in an emptyDataFrame
(GH 19157)Bug in
Timestamp.tz_localize()
where localizing a timestamp near the minimum or maximum valid values could overflow and return a timestamp with an incorrect nanosecond value (GH 12677)Bug when iterating over
DatetimeIndex
that was localized with fixed timezone offset that rounded nanosecond precision to microseconds (GH 19603)Bug in
DataFrame.diff()
that raised anIndexError
with tzaware values (GH 18578)Bug in
melt()
that converted tzaware dtypes to tznaive (GH 15785)Bug in
Dataframe.count()
that raised anValueError
, ifDataframe.dropna()
was called for a single column with timezoneaware values. (GH 13407)
Offsets#
Bug in
WeekOfMonth
andWeek
where addition and subtraction did not roll correctly (GH 18510, GH 18672, GH 18864)Bug in
WeekOfMonth
andLastWeekOfMonth
where default keyword arguments for constructor raisedValueError
(GH 19142)Bug in
FY5253Quarter
,LastWeekOfMonth
where rollback and rollforward behavior was inconsistent with addition and subtraction behavior (GH 18854)Bug in
FY5253
wheredatetime
addition and subtraction incremented incorrectly for dates on the yearend but not normalized to midnight (GH 18854)Bug in
FY5253
where date offsets could incorrectly raise anAssertionError
in arithmetic operations (GH 14774)
Numeric#
Bug in
Series
constructor with an int or float list where specifyingdtype=str
,dtype='str'
ordtype='U'
failed to convert the data elements to strings (GH 16605)Bug in
Index
multiplication and division methods where operating with aSeries
would return anIndex
object instead of aSeries
object (GH 19042)Bug in the
DataFrame
constructor in which data containing very large positive or very large negative numbers was causingOverflowError
(GH 18584)Bug in
Index
constructor withdtype='uint64'
where intlike floats were not coerced toUInt64Index
(GH 18400)Bug in
DataFrame
flex arithmetic (e.g.df.add(other, fill_value=foo)
) with afill_value
other thanNone
failed to raiseNotImplementedError
in corner cases where either the frame orother
has length zero (GH 19522)Multiplication and division of numericdtyped
Index
objects with timedeltalike scalars returnsTimedeltaIndex
instead of raisingTypeError
(GH 19333)Bug where
NaN
was returned instead of 0 bySeries.pct_change()
andDataFrame.pct_change()
whenfill_method
is notNone
(GH 19873)
Strings#
Bug in
Series.str.get()
with a dictionary in the values and the index not in the keys, raisingKeyError
(GH 20671)
Indexing#
Bug in
Index
construction from list of mixed type tuples (GH 18505)Bug in
Index.drop()
when passing a list of both tuples and nontuples (GH 18304)Bug in
DataFrame.drop()
,Panel.drop()
,Series.drop()
,Index.drop()
where noKeyError
is raised when dropping a nonexistent element from an axis that contains duplicates (GH 19186)Bug in indexing a datetimelike
Index
that raisedValueError
instead ofIndexError
(GH 18386).Index.to_series()
now acceptsindex
andname
kwargs (GH 18699)DatetimeIndex.to_series()
now acceptsindex
andname
kwargs (GH 18699)Bug in indexing nonscalar value from
Series
having nonuniqueIndex
will return value flattened (GH 17610)Bug in indexing with iterator containing only missing keys, which raised no error (GH 20748)
Fixed inconsistency in
.ix
between list and scalar keys when the index has integer dtype and does not include the desired keys (GH 20753)Bug in
__setitem__
when indexing aDataFrame
with a 2d boolean ndarray (GH 18582)Bug in
str.extractall
when there were no matches emptyIndex
was returned instead of appropriateMultiIndex
(GH 19034)Bug in
IntervalIndex
where empty and purely NA data was constructed inconsistently depending on the construction method (GH 18421)Bug in
IntervalIndex.symmetric_difference()
where the symmetric difference with a nonIntervalIndex
did not raise (GH 18475)Bug in
IntervalIndex
where set operations that returned an emptyIntervalIndex
had the wrong dtype (GH 19101)Bug in
DataFrame.drop_duplicates()
where noKeyError
is raised when passing in columns that don’t exist on theDataFrame
(GH 19726)Bug in
Index
subclasses constructors that ignore unexpected keyword arguments (GH 19348)Bug in
Index.difference()
when taking difference of anIndex
with itself (GH 20040)Bug in
DataFrame.first_valid_index()
andDataFrame.last_valid_index()
in presence of entire rows of NaNs in the middle of values (GH 20499).Bug in
IntervalIndex
where some indexing operations were not supported for overlapping or nonmonotonicuint64
data (GH 20636)Bug in
Series.is_unique
where extraneous output in stderr is shown if Series contains objects with__ne__
defined (GH 20661)Bug in
.loc
assignment with a singleelement listlike incorrectly assigns as a list (GH 19474)Bug in partial string indexing on a
Series/DataFrame
with a monotonic decreasingDatetimeIndex
(GH 19362)Bug in performing inplace operations on a
DataFrame
with a duplicateIndex
(GH 17105)Bug in
IntervalIndex.get_loc()
andIntervalIndex.get_indexer()
when used with anIntervalIndex
containing a single interval (GH 17284, GH 20921)Bug in
.loc
with auint64
indexer (GH 20722)
MultiIndex#
Bug in
MultiIndex.__contains__()
where nontuple keys would returnTrue
even if they had been dropped (GH 19027)Bug in
MultiIndex.set_labels()
which would cause casting (and potentially clipping) of the new labels if thelevel
argument is not 0 or a list like [0, 1, … ] (GH 19057)Bug in
MultiIndex.get_level_values()
which would return an invalid index on level of ints with missing values (GH 17924)Bug in
MultiIndex.unique()
when called on emptyMultiIndex
(GH 20568)Bug in
MultiIndex.unique()
which would not preserve level names (GH 20570)Bug in
MultiIndex.remove_unused_levels()
which would fill nan values (GH 18417)Bug in
MultiIndex.from_tuples()
which would fail to take zipped tuples in python3 (GH 18434)Bug in
MultiIndex.get_loc()
which would fail to automatically cast values between float and int (GH 18818, GH 15994)Bug in
MultiIndex.get_loc()
which would cast boolean to integer labels (GH 19086)Bug in
MultiIndex.get_loc()
which would fail to locate keys containingNaN
(GH 18485)Bug in
MultiIndex.get_loc()
in largeMultiIndex
, would fail when levels had different dtypes (GH 18520)Bug in indexing where nested indexers having only numpy arrays are handled incorrectly (GH 19686)
IO#
read_html()
now rewinds seekable IO objects after parse failure, before attempting to parse with a new parser. If a parser errors and the object is nonseekable, an informative error is raised suggesting the use of a different parser (GH 17975)DataFrame.to_html()
now has an option to add an id to the leading<table>
tag (GH 8496)Bug in
read_msgpack()
with a non existent file is passed in Python 2 (GH 15296)Bug in
read_csv()
where aMultiIndex
with duplicate columns was not being mangled appropriately (GH 18062)Bug in
read_csv()
where missing values were not being handled properly whenkeep_default_na=False
with dictionaryna_values
(GH 19227)Bug in
read_csv()
causing heap corruption on 32bit, bigendian architectures (GH 20785)Bug in
read_sas()
where a file with 0 variables gave anAttributeError
incorrectly. Now it gives anEmptyDataError
(GH 18184)Bug in
DataFrame.to_latex()
where pairs of braces meant to serve as invisible placeholders were escaped (GH 18667)Bug in
DataFrame.to_latex()
where aNaN
in aMultiIndex
would cause anIndexError
or incorrect output (GH 14249)Bug in
DataFrame.to_latex()
where a nonstring indexlevel name would result in anAttributeError
(GH 19981)Bug in
DataFrame.to_latex()
where the combination of an index name and theindex_names=False
option would result in incorrect output (GH 18326)Bug in
DataFrame.to_latex()
where aMultiIndex
with an empty string as its name would result in incorrect output (GH 18669)Bug in
DataFrame.to_latex()
where missing space characters caused wrong escaping and produced nonvalid latex in some cases (GH 20859)Bug in
read_json()
where large numeric values were causing anOverflowError
(GH 18842)Bug in
DataFrame.to_parquet()
where an exception was raised if the write destination is S3 (GH 19134)Interval
now supported inDataFrame.to_excel()
for all Excel file types (GH 19242)Timedelta
now supported inDataFrame.to_excel()
for all Excel file types (GH 19242, GH 9155, GH 19900)Bug in
pandas.io.stata.StataReader.value_labels()
raising anAttributeError
when called on very old files. Now returns an empty dict (GH 19417)Bug in
read_pickle()
when unpickling objects withTimedeltaIndex
orFloat64Index
created with pandas prior to version 0.20 (GH 19939)Bug in
pandas.io.json.json_normalize()
where subrecords are not properly normalized if any subrecords values are NoneType (GH 20030)Bug in
usecols
parameter inread_csv()
where error is not raised correctly when passing a string. (GH 20529)Bug in
HDFStore.keys()
when reading a file with a soft link causes exception (GH 20523)Bug in
HDFStore.select_column()
where a key which is not a valid store raised anAttributeError
instead of aKeyError
(GH 17912)
Plotting#
Better error message when attempting to plot but matplotlib is not installed (GH 19810).
DataFrame.plot()
now raises aValueError
when thex
ory
argument is improperly formed (GH 18671)Bug in
DataFrame.plot()
whenx
andy
arguments given as positions caused incorrect referenced columns for line, bar and area plots (GH 20056)Bug in formatting tick labels with
datetime.time()
and fractional seconds (GH 18478).Series.plot.kde()
has exposed the argsind
andbw_method
in the docstring (GH 18461). The argumentind
may now also be an integer (number of sample points).DataFrame.plot()
now supports multiple columns to they
argument (GH 19699)
GroupBy/resample/rolling#
Bug when grouping by a single column and aggregating with a class like
list
ortuple
(GH 18079)Fixed regression in
DataFrame.groupby()
which would not emit an error when called with a tuple key not in the index (GH 18798)Bug in
DataFrame.resample()
which silently ignored unsupported (or mistyped) options forlabel
,closed
andconvention
(GH 19303)Bug in
DataFrame.groupby()
where tuples were interpreted as lists of keys rather than as keys (GH 17979, GH 18249)Bug in
DataFrame.groupby()
where aggregation byfirst
/last
/min
/max
was causing timestamps to lose precision (GH 19526)Bug in
DataFrame.transform()
where particular aggregation functions were being incorrectly cast to match the dtype(s) of the grouped data (GH 19200)Bug in
DataFrame.groupby()
passing theon=
kwarg, and subsequently using.apply()
(GH 17813)Bug in
DataFrame.resample().aggregate
not raising aKeyError
when aggregating a nonexistent column (GH 16766, GH 19566)Bug in
DataFrameGroupBy.cumsum()
andDataFrameGroupBy.cumprod()
whenskipna
was passed (GH 19806)Bug in
DataFrame.resample()
that dropped timezone information (GH 13238)Bug in
DataFrame.groupby()
where transformations usingnp.all
andnp.any
were raising aValueError
(GH 20653)Bug in
DataFrame.resample()
whereffill
,bfill
,pad
,backfill
,fillna
,interpolate
, andasfreq
were ignoringloffset
. (GH 20744)Bug in
DataFrame.groupby()
when applying a function that has mixed data types and the user supplied function can fail on the grouping column (GH 20949)Bug in
DataFrameGroupBy.rolling().apply()
where operations performed against the associatedDataFrameGroupBy
object could impact the inclusion of the grouped item(s) in the result (GH 14013)
Sparse#
Bug in which creating a
SparseDataFrame
from a denseSeries
or an unsupported type raised an uncontrolled exception (GH 19374)Bug in
SparseDataFrame.to_csv
causing exception (GH 19384)Bug in
SparseSeries.memory_usage
which caused segfault by accessing non sparse elements (GH 19368)Bug in constructing a
SparseArray
: ifdata
is a scalar andindex
is defined it will coerce tofloat64
regardless of scalar’s dtype. (GH 19163)
Reshaping#
Bug in
DataFrame.merge()
where referencing aCategoricalIndex
by name, where theby
kwarg wouldKeyError
(GH 20777)Bug in
DataFrame.stack()
which fails trying to sort mixed type levels under Python 3 (GH 18310)Bug in
DataFrame.unstack()
which casts int to float ifcolumns
is aMultiIndex
with unused levels (GH 17845)Bug in
DataFrame.unstack()
which raises an error ifindex
is aMultiIndex
with unused labels on the unstacked level (GH 18562)Fixed construction of a
Series
from adict
containingNaN
as key (GH 18480)Fixed construction of a
DataFrame
from adict
containingNaN
as key (GH 18455)Disabled construction of a
Series
where len(index) > len(data) = 1, which previously would broadcast the data item, and now raises aValueError
(GH 18819)Suppressed error in the construction of a
DataFrame
from adict
containing scalar values when the corresponding keys are not included in the passed index (GH 18600)Fixed (changed from
object
tofloat64
) dtype ofDataFrame
initialized with axes, no data, anddtype=int
(GH 19646)Bug in
Series.rank()
whereSeries
containingNaT
modifies theSeries
inplace (GH 18521)Bug in
cut()
which fails when using readonly arrays (GH 18773)Bug in
DataFrame.pivot_table()
which fails when theaggfunc
arg is of type string. The behavior is now consistent with other methods likeagg
andapply
(GH 18713)Bug in
DataFrame.merge()
in which merging usingIndex
objects as vectors raised an Exception (GH 19038)Bug in
DataFrame.stack()
,DataFrame.unstack()
,Series.unstack()
which were not returning subclasses (GH 15563)Bug in timezone comparisons, manifesting as a conversion of the index to UTC in
.concat()
(GH 18523)Bug in
concat()
when concatenating sparse and dense series it returns only aSparseDataFrame
. Should be aDataFrame
. (GH 18914, GH 18686, and GH 16874)Improved error message for
DataFrame.merge()
when there is no common merge key (GH 19427)Bug in
DataFrame.join()
which does anouter
instead of aleft
join when being called with multiple DataFrames and some have nonunique indices (GH 19624)Series.rename()
now acceptsaxis
as a kwarg (GH 18589)Bug in
rename()
where an Index of samelength tuples was converted to a MultiIndex (GH 19497)Comparisons between
Series
andIndex
would return aSeries
with an incorrect name, ignoring theIndex
’s name attribute (GH 19582)Bug in
qcut()
where datetime and timedelta data withNaT
present raised aValueError
(GH 19768)Bug in
DataFrame.iterrows()
, which would infers strings not compliant to ISO8601 to datetimes (GH 19671)Bug in
Series
constructor withCategorical
where aValueError
is not raised when an index of different length is given (GH 19342)Bug in
DataFrame.astype()
where column metadata is lost when converting to categorical or a dictionary of dtypes (GH 19920)Bug in
cut()
andqcut()
where timezone information was dropped (GH 19872)Bug in
Series
constructor with adtype=str
, previously raised in some cases (GH 19853)Bug in
get_dummies()
, andselect_dtypes()
, where duplicate column names caused incorrect behavior (GH 20848)Bug in
isna()
, which cannot handle ambiguous typed lists (GH 20675)Bug in
concat()
which raises an error when concatenating TZaware dataframes and allNaT dataframes (GH 12396)Bug in
concat()
which raises an error when concatenating empty TZaware series (GH 18447)
Other#
Improved error message when attempting to use a Python keyword as an identifier in a
numexpr
backed query (GH 18221)Bug in accessing a
pandas.get_option()
, which raisedKeyError
rather thanOptionError
when looking up a nonexistent option key in some cases (GH 19789)Bug in
testing.assert_series_equal()
andtesting.assert_frame_equal()
for Series or DataFrames with differing unicode data (GH 20503)
Contributors#
A total of 328 people contributed patches to this release. People with a “+” by their names contributed a patch for the first time.
Aaron Critchley
AbdealiJK +
Adam Hooper +
Albert Villanova del Moral
Alejandro Giacometti +
Alejandro Hohmann +
Alex Rychyk
Alexander Buchkovsky
Alexander Lenail +
Alexander Michael Schade
Aly Sivji +
Andreas Költringer +
Andrew
Andrew Bui +
András Novoszáth +
Andy Craze +
Andy R. Terrel
Anh Le +
Anil Kumar Pallekonda +
Antoine Pitrou +
Antonio Linde +
Antonio Molina +
Antonio Quinonez +
Armin Varshokar +
Artem Bogachev +
Avi Sen +
Azeez Oluwafemi +
Ben Auffarth +
Bernhard Thiel +
Bhavesh Poddar +
BielStela +
Blair +
Bob Haffner
Brett Naul +
Brock Mendel
Bryce Guinta +
Carlos Eduardo Moreira dos Santos +
Carlos García Márquez +
Carol Willing
Cheuk Ting Ho +
Chitrank Dixit +
Chris
Chris Burr +
Chris Catalfo +
Chris Mazzullo
Christian Chwala +
Cihan Ceyhan +
Clemens Brunner
Colin +
Cornelius Riemenschneider
Crystal Gong +
DaanVanHauwermeiren
Dan Dixey +
Daniel Frank +
Daniel Garrido +
Daniel Sakuma +
DataOmbudsman +
Dave Hirschfeld
Dave Lewis +
David Adrián Cañones Castellano +
David Arcos +
David C Hall +
David Fischer
David Hoese +
David Lutz +
David Polo +
David Stansby
Dennis Kamau +
Dillon Niederhut
Dimitri +
Dr. Irv
Dror Atariah
Eric Chea +
Eric Kisslinger
Eric O. LEBIGOT (EOL) +
FANGOD +
Fabian Retkowski +
Fer Sar +
Gabriel de Maeztu +
Gianpaolo Macario +
Giftlin Rajaiah
Gilberto Olimpio +
Gina +
Gjelt +
Graham Inggs +
Grant Roch
Grant Smith +
Grzegorz Konefał +
Guilherme Beltramini
HagaiHargil +
Hamish Pitkeathly +
Hammad Mashkoor +
Hannah Ferchland +
Hans
Haochen Wu +
Hissashi Rocha +
Iain Barr +
Ibrahim Sharaf ElDen +
Ignasi Fosch +
Igor Conrado Alves de Lima +
Igor Shelvinskyi +
Imanflow +
Ingolf Becker
Israel Saeta Pérez
Iva Koevska +
Jakub Nowacki +
Jan FF +
Jan Koch +
Jan Werkmann
Janelle Zoutkamp +
Jason Bandlow +
Jaume Bonet +
Jay Alammar +
Jeff Reback
JennaVergeynst
Jimmy Woo +
Jing Qiang Goh +
Joachim Wagner +
Joan Martin Miralles +
Joel Nothman
Joeun Park +
John Cant +
Johnny Metz +
Jon Mease
Jonas Schulze +
Jongwony +
Jordi Contestí +
Joris Van den Bossche
José F. R. Fonseca +
Jovixe +
Julio Martinez +
Jörg Döpfert
KOBAYASHI Ittoku +
Kate Surta +
Kenneth +
Kevin Kuhl
Kevin Sheppard
Krzysztof Chomski
Ksenia +
Ksenia Bobrova +
Kunal Gosar +
Kurtis Kerstein +
Kyle Barron +
Laksh Arora +
Laurens Geffert +
Leif Walsh
Liam Marshall +
Liam3851 +
Licht Takeuchi
Liudmila +
Ludovico Russo +
Mabel Villalba +
Manan Pal Singh +
Manraj Singh
Marc +
Marc Garcia
Marco Hemken +
Maria del Mar Bibiloni +
Mario Corchero +
Mark Woodbridge +
Martin Journois +
Mason Gallo +
Matias Heikkilä +
Matt BraymerHayes
Matt Kirk +
Matt Maybeno +
Matthew Kirk +
Matthew Rocklin +
Matthew Roeschke
Matthias Bussonnier +
Max Mikhaylov +
Maxim Veksler +
Maximilian Roos
Maximiliano Greco +
Michael Penkov
Michael Röttger +
Michael Selik +
Michael Waskom
Mie~~~
Mike Kutzma +
Ming Li +
Mitar +
Mitch Negus +
Montana Low +
Moritz Münst +
Mortada Mehyar
Myles Braithwaite +
Nate Yoder
Nicholas Ursa +
Nick Chmura
Nikos Karagiannakis +
Nipun Sadvilkar +
Nis Martensen +
Noah +
Noémi Éltető +
Olivier Bilodeau +
Ondrej Kokes +
Onno Eberhard +
Paul Ganssle +
Paul Mannino +
Paul Reidy
Paulo Roberto de Oliveira Castro +
Pepe Flores +
Peter Hoffmann
Phil Ngo +
Pietro Battiston
Pranav Suri +
Priyanka Ojha +
Pulkit Maloo +
README Bot +
Ray Bell +
Riccardo Magliocchetti +
Ridhwan Luthra +
Robert Meyer
Robin
Robin Kiplang’at +
Rohan Pandit +
Rok Mihevc +
Rouz Azari
Ryszard T. Kaleta +
Sam Cohan
Sam Foo
Samir Musali +
Samuel Sinayoko +
Sangwoong Yoon
SarahJessica +
Sharad Vijalapuram +
Shubham Chaudhary +
SiYoungOh +
Sietse Brouwer
Simone Basso +
Stefania Delprete +
Stefano Cianciulli +
Stephen Childs +
StephenVoland +
Stijn Van Hoey +
Sven
Talitha Pumar +
Tarbo Fukazawa +
Ted Petrou +
Thomas A Caswell
Tim Hoffmann +
Tim Swast
Tom Augspurger
Tommy +
Tulio Casagrande +
Tushar Gupta +
Tushar Mittal +
Upkar Lidder +
Victor Villas +
Vince W +
Vinícius Figueiredo +
Vipin Kumar +
WBare
Wenhuan +
Wes Turner
William Ayd
Wilson Lin +
Xbar
Yaroslav Halchenko
Yee Mey
Yeongseon Choe +
Yian +
Yimeng Zhang
ZhuBaohe +
Zihao Zhao +
adatasetaday +
akielbowicz +
akosel +
alinde1 +
amuta +
bolkedebruin
cbertinato
cgohlke
charlie0389 +
chrisb1
csfarkas +
dajcs +
deflatSOCO +
derestlehtwg
discort
dmanikowskireef +
donK23 +
elrubio +
fivemok +
fjdiod
fjetter +
froessler +
gabrielclow
gfyoung
ghasemnaddaf
hvetinari +
himanshu awasthi +
ignamv +
jayfoad +
jazzmuesli +
jbrockmendel
jen w +
jjames34 +
joaoavf +
joders +
jschendel
juan huguet +
l736x +
luzpaz +
mdeboc +
miguelmorin +
miker985
miquelcamprodon +
orereta +
ottiP +
peterpanmj +
rafarui +
raphm +
readyready15728 +
rmihael +
samghelms +
scriptomation +
sfoo +
stefansimik +
stonebig
tmnhat2001 +
tomneep +
topper123
tv3141 +
verakai +
xpvpc +
zhanghui +