DOC: read_csv: date_format

[igurin-invn]

Pandas version checks

• I have checked that the issue still exists on the latest versions of the docs on main here

Location of the documentation

https://pandas.pydata.org/docs/dev/reference/api/pandas.read_csv.html

Documentation problem

When using read_csv, I came across a UserWarning: Could not infer format.... I tried to resolve it by providing a date_format, but the warning wouldn't go away. Then I realized that my problem was that I was providing specifically the date format but not the time format.

The documentation for this parameter is confusing. It starts with what appear to be two alternate and contradictory definitions of the parameter:

• "Format to use for parsing dates when used in conjunction with parse_dates."
• "The strftime to parse time, e.g. "%d/%m/%Y"."

Suggested fix for documentation

Say something like this:

"Format to use for parsing dates and/or times. Used in conjunction with parse_dates. See strftime documentation for more information on choices..."

[rhshadrach]

Thanks for the report. Does this always result in a datetime, or can it result in a date and/or time? I think that would determine whether the documentation should be refined to say datetime or date/time/datetime.

It looks like someone didn't proofread enough.

Reports and contributions are welcome, this is not.

[igurin-invn]

Does this always result in a datetime, or can it result in a date and/or time? I think that would determine whether the documentation should be refined to say datetime or date/time/datetime.

Looks like the result is always a DatetimeIndex.

>>> from io import StringIO
>>> import pandas as pd
>>> pd.read_csv(StringIO(',A\n7/12/2024 5:40:41 PM,0\n7/12/2024 5:41:41 PM,0'), # date and time
        parse_dates=True, index_col=0).index
DatetimeIndex(['2024-07-12 17:40:41', '2024-07-12 17:41:41'], dtype='datetime64[ns]', freq=None)
>>> pd.read_csv(StringIO(',A\n7/12/2024,0\n7/12/2024,0'),                 # date only
        parse_dates=True, index_col=0).index
DatetimeIndex(['2024-07-12', '2024-07-12'], dtype='datetime64[ns]', freq=None)
>>> pd.read_csv(StringIO(',A\n5:40:41 PM,0\n5:41:41 PM,0'),               # time only
        parse_dates=True, index_col=0).index
DatetimeIndex(['2024-08-20 17:40:41', '2024-08-20 17:41:41'], dtype='datetime64[ns]', freq=None)

Reports and contributions are welcome, this is not.

Sorry. No offense intended. I took it out.