Changing certain column type to date from datetime

January 11, 2022

I have a df in which I want to change certain columns type to date from datetime:

field   category    2022-01-10 00:00:00     2022-01-17 00:00:00     2022-01-24 00:00:00
A       10          500                     700                     500
B       15          60                      70                      50

I am trying to achieve this:

field   category    2022-01-10      2022-01-17      2022-01-24
A       10          500             700             500
B       15          60              70              50

Because during concat with other df the column type changes from 2021-11-19 to 2021-11-19 00:00:00.

From this answer I tried::

df.loc[:,pd.to_datetime('2022-01-10 00:00:00') : pd.to_datetime('2022-06-20 00:00:00')].columns = 
                                                 pd.to_datetime(df.loc[:,pd.to_datetime('2022-01-10 00:00:00') : 
                                                 pd.to_datetime('2022-06-20 00:00:00')].columns)
                                                 .strftime('%Y-%m-%d')

But this does not change the type of the column which I cant understand why because

pd.to_datetime(df.loc[:,pd.to_datetime('2022-01-10 00:00:00') : pd.to_datetime('2022-06-20 00:00:00')].columns).strftime('%Y-%m-%d')

Return:

Index(['2022-01-10', '2022-01-17', '2022-01-24')]

>Solution :

In your column index you mix string (‘field’ and ‘category’) and timestamp (other columns) so your column index is an Index not a DatetimeIndex.

>>> df.columns
Index([            'field',          'category', 2022-01-10 00:00:00,
       2022-01-17 00:00:00, 2022-01-24 00:00:00],
      dtype='object')

>>> [type(c) for c in df.columns]
[str,
 str,
 pandas._libs.tslibs.timestamps.Timestamp,
 pandas._libs.tslibs.timestamps.Timestamp,
 pandas._libs.tslibs.timestamps.Timestamp]

The default string representation of Timestamp is YYYY-mm-dd HH:MM:SS while the default representation of each value of a DatetimeIndex can be YYYY-mm-dd only (if possible).

To get the date format, you have to cast your Timestamp to string.