What’s new in 1.4.4 (August 31, 2022)

These are the changes in pandas 1.4.4. See Release notes for a full changelog including other versions of pandas.

Fixed regressions

• Fixed regression in DataFrame.fillna() not working on a DataFrame with a MultiIndex (GH 47649)

• Fixed regression in taking NULL objects from a DataFrame causing a segmentation violation. These NULL values are created by numpy.empty_like() (GH 46848)

• Fixed regression in concat() materializing the Index during sorting even if the Index was already sorted (GH 47501)

• Fixed regression in concat() or merge() handling of all-NaN ExtensionArrays with custom attributes (GH 47762)

• Fixed regression in calling bitwise numpy ufuncs (for example, np.bitwise_and) on Index objects (GH 46769)

• Fixed regression in cut() when using a datetime64 IntervalIndex as bins (GH 46218)

• Fixed regression in DataFrame.select_dtypes() where include="number" included BooleanDtype (GH 46870)

• Fixed regression in DataFrame.loc() raising error when indexing with a NamedTuple (GH 48124)

• Fixed regression in DataFrame.loc() not updating the cache correctly after values were set (GH 47867)

• Fixed regression in DataFrame.loc() not aligning index in some cases when setting a DataFrame (GH 47578)

• Fixed regression in DataFrame.loc() setting a length-1 array like value to a single value in the DataFrame (GH 46268)

• Fixed regression when slicing with DataFrame.loc() with DatetimeIndex with a DateOffset object for its freq (GH 46671)

• Fixed regression in setting None or non-string value into a string-dtype Series using a mask (GH 47628)

• Fixed regression in updating a DataFrame column through Series __setitem__ (using chained assignment) not updating column values inplace and using too much memory (GH 47172)

• Fixed regression in DataFrame.select_dtypes() returning a view on the original DataFrame (GH 48090)

• Fixed regression using custom Index subclasses (for example, used in xarray) with reset_index() or Index.insert() (GH 47071)

• Fixed regression in intersection() when the DatetimeIndex has dates crossing daylight savings time (GH 46702)

• Fixed regression in merge() throwing an error when passing a Series with a multi-level name (GH 47946)

• Fixed regression in DataFrame.eval() creating a copy when updating inplace (GH 47449)

• Fixed regression where getting a row using DataFrame.iloc() with SparseDtype would raise (GH 46406)

Bug fixes

• The FutureWarning raised when passing arguments (other than filepath_or_buffer) as positional in read_csv() is now raised at the correct stacklevel (GH 47385)

• Bug in DataFrame.to_sql() when method was a callable that did not return an int and would raise a TypeError (GH 46891)

• Bug in DataFrameGroupBy.value_counts() where subset had no effect (GH 46383)

• Bug when getting values with DataFrame.loc() with a list of keys causing an internal inconsistency that could lead to a disconnect between frame.at[x, y] vs frame[y].loc[x] (GH 22372)

• Bug in the Series.dt.strftime() accessor return a float instead of object dtype Series for all-NaT input, which also causes a spurious deprecation warning (GH 45858)

Other

• The minimum version of Cython needed to compile pandas is now 0.29.32 (GH 47978)

Contributors

A total of 12 people contributed patches to this release. People with a “+” by their names contributed a patch for the first time.

• Jeff Reback

• Joris Van den Bossche

• Marco Edward Gorelli

• Matthew Roeschke

• MeeseeksMachine

• Nima Sarang +

• Pandas Development Team

• Patrick Hoefler

• Simon Hawkins

• Tobias Pitters

• Xingrong Chen +

• jbrockmendel