Skip to content

fillna() on Series or DataFrame containing datetime64 mess the values #6587

Closed
@yonil7

Description

@yonil7

s = pd.Series([pd.NaT, pd.NaT, '2013-08-05 15:30:00.000001'])
print s
0 NaT
1 NaT
2 2013-08-05 15:30:00.000001
dtype: datetime64[ns]

print s.fillna(method='backfill')
0 2013-08-05 15:30:00.000001024
1 2013-08-05 15:30:00.000001024
2 2013-08-05 15:30:00.000001024
dtype: datetime64[ns]

This happens with pandas 0.13.1 and numpy 1.8.0

Metadata

Metadata

Assignees

No one assigned

    Labels

    BugDatetimeDatetime data dtypeDtype ConversionsUnexpected or buggy dtype conversionsMissing-datanp.nan, pd.NaT, pd.NA, dropna, isnull, interpolate

    Type

    No type

    Projects

    No projects

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions