Pandas：astype datetime 分钟失败

Question

I am using pandas 1.5.2.

I have a column of timestamp stored in a pandas dataframe.

wall['Timestamp'].astype('datetime64[Y]')

0    2017-07-11 09:33:14.819
1    2017-07-11 09:33:14.819
2    2017-07-11 09:33:14.819
3    2017-07-11 09:33:14.819
4    2017-07-11 09:33:14.820
         
5   2017-07-11 16:20:52.463
6   2017-07-11 16:20:52.463
7   2017-07-11 16:20:52.463
8   2017-07-11 16:20:52.464
9   2017-07-11 16:20:54.984

I used to be able to convert the timestamp to '%Y-%m-%d %H:%M' by df['Timestamp'].astype('datetime64[m]') . But now this doesn't work and returns the original column.

0    2017-07-11 09:33
1    2017-07-11 09:33
2    2017-07-11 09:33
3    2017-07-11 09:33
4    2017-07-11 09:33
         
5   2017-07-11 16:20
6   2017-07-11 16:20
7   2017-07-11 16:20
8   2017-07-11 16:20
9   2017-07-11 16:20

Answer 1

You can change a format of the column like this:

import pandas as pd

df = pd.DataFrame({
    'Timestamp': ['2017-07-11 09:33:14.819', '2017-07-11 09:33:14.819']
})
print(df)
"""
                 Timestamp
0  2017-07-11 09:33:14.819
1  2017-07-11 09:33:14.819
"""

df['Timestamp'] = df['Timestamp'].astype('datetime64[m]')
print(df)
"""
            Timestamp
0 2017-07-11 09:33:00
1 2017-07-11 09:33:00
"""

If you want to remove the second format, you can convert the value to str and get only the part you want using slicing , as follows:

df['Timestamp'] = df['Timestamp'].astype(str).str[0:16]
print(df)
"""
          Timestamp
0  2017-07-11 09:33
1  2017-07-11 09:33
"""

Answer 2

I guess you can use the strftime or to_datetime package.

wall['Timestamp'] = wall['Timestamp'].dt.strftime('%Y-%m-%d %H:%M')

Or

wall['Timestamp'] = pd.to_datetime(wall['Timestamp']).dt.strftime('%Y-%m-%d %H:%M')