What’s new in 1.1.2 (September 8, 2020)¶
These are the changes in pandas 1.1.2. See Release notes for a full changelog including other versions of pandas.
Fixed regressions¶
Regression in
DatetimeIndex.intersection()
incorrectly raisingAssertionError
when intersecting against a list (GH35876)Fix regression in updating a column inplace (e.g. using
df['col'].fillna(.., inplace=True)
) (GH35731)Fix regression in
DataFrame.append()
mixing tz-aware and tz-naive datetime columns (GH35460)Performance regression for
RangeIndex.format()
(GH35712)Regression where
MultiIndex.get_loc()
would return a slice spanning the full index when passed an empty list (GH35878)Fix regression in invalid cache after an indexing operation; this can manifest when setting which does not update the data (GH35521)
Regression in
DataFrame.replace()
where aTypeError
would be raised when attempting to replace elements of typeInterval
(GH35931)Fix regression in pickle roundtrip of the
closed
attribute ofIntervalIndex
(GH35658)Fixed regression in
DataFrameGroupBy.agg()
where aValueError: buffer source array is read-only
would be raised when the underlying array is read-only (GH36014)Fixed regression in
Series.groupby.rolling()
number of levels ofMultiIndex
in input was compressed to one (GH36018)Fixed regression in
DataFrameGroupBy
on an emptyDataFrame
(GH36197)
Bug fixes¶
Bug in
DataFrame.eval()
withobject
dtype column binary operations (GH35794)Bug in
Series
constructor raising aTypeError
when constructing sparse datetime64 dtypes (GH35762)Bug in
DataFrame.apply()
withresult_type="reduce"
returning with incorrect index (GH35683)Bug in
Series.astype()
andDataFrame.astype()
not respecting theerrors
argument when set to"ignore"
for extension dtypes (GH35471)Bug in
DateTimeIndex.format()
andPeriodIndex.format()
withname=True
setting the first item to"None"
where it should be""
(GH35712)Bug in
Float64Index.__contains__()
incorrectly raisingTypeError
instead of returningFalse
(GH35788)Bug in
Series
constructor incorrectly raising aTypeError
when passed an ordered set (GH36044)Bug in
Series.dt.isocalendar()
andDatetimeIndex.isocalendar()
that returned incorrect year for certain dates (GH36032)Bug in
DataFrame
indexing returning an incorrectSeries
in some cases when the series has been altered and a cache not invalidated (GH33675)Bug in
DataFrame.corr()
causing subsequent indexing lookups to be incorrect (GH35882)Bug in
import_optional_dependency()
returning incorrect package names in cases where package name is different from import name (GH35948)Bug when setting empty
DataFrame
column to aSeries
in preserving name of index in frame (GH31368)
Other¶
factorize()
now supportsna_sentinel=None
to include NaN in the uniques of the values and removedropna
keyword which was unintentionally exposed to public facing API in 1.1 version fromfactorize()
(GH35667)DataFrame.plot()
andSeries.plot()
raiseUserWarning
about usage ofFixedFormatter
andFixedLocator
(GH35684 and GH35945)