Warning
Starting with the 0.25.x series of releases, pandas only supports Python 3.5 and higher. See install.dropping-27
for more details.
{{ header }}
These are the changes in pandas 0.25.0. See release
for a full changelog including other versions of pandas.
Timestamp.replace
now supports thefold
argument to disambiguate DST transition times (25017
)
DatetimeTZDtype
will now standardize pytz timezones to a common timezone instance (24713
)
- Deprecated the M (months) and Y (year) units parameter of :func: pandas.to_timedelta, :func: pandas.Timedelta and :func: pandas.TimedeltaIndex (
16344
)
- Removed (parts of)
Panel
(25047
,25191
,25231
)
- Significant speedup in SparseArray initialization that benefits most operations, fixing performance regression introduced in v0.20.0 (
24985
) - DataFrame.to_stata() is now faster when outputting data with any string or non-native endian columns (
25045
)
- Bug in
to_datetime
withutc=True
and datetime strings that would apply previously parsed UTC offsets to subsequent arguments (24992
) - Bug in
Timestamp.tz_localize
andTimestamp.tz_convert
does not propagatefreq
(25241
)
- Bug in
to_numeric
in which large negative numbers were being improperly handled (24910
) - Bug in
to_numeric
in which numbers were being coerced to float, even thougherrors
was notcoerce
(24910
)
- Fixed misleading exception message in
Series.missing
if argumentorder
is required, but omitted (10633
,24014
).
- Fixed bug in missing text when using
to_clipboard
if copying utf-16 characters in Python 3 on Windows (25040
)
- Bug in
pandas.merge
adds a string ofNone
ifNone
is assigned in suffixes instead of remain the column name as-is (24782
). - Bug in
merge
when merging by index name would sometimes result in an incorrectly numbered index (24212
) to_records
now accepts dtypes to its column_dtypes parameter (24895
)
- Significant speedup in SparseArray initialization that benefits most operations, fixing performance regression introduced in v0.20.0 (
24985
)
v0.24.x..HEAD