pd datetime

Basic dataframe

import pandas as pd
import numpy as np
dates = [str(y)+"-" +str(m)+"-"+str(d) for y in range(2000, 2001) for m in range(1, 13) for d in range(1, 20)]
df = pd.DataFrame({"date": dates})

Change string to datetime

>>> df["date_datetime"] = pd.to_datetime(df["date"])
>>> df.dtypes
date                     object
date_datetime    datetime64[ns]

Set a missing value

>>> df.iloc[0, 1] = np.NaN
>>> df
           date date_datetime
0      2000-1-1           NaT
1      2000-1-2    2000-01-02
2      2000-1-3    2000-01-03
3      2000-1-4    2000-01-04
4      2000-1-5    2000-01-05
..          ...           ...
223  2000-12-15    2000-12-15
224  2000-12-16    2000-12-16
225  2000-12-17    2000-12-17
226  2000-12-18    2000-12-18
227  2000-12-19    2000-12-19

Change datetime to the first day of a month [1]

>>> df["year_month_firstDay"] = df['date_datetime'].to_numpy().astype('datetime64[M]')
>>> df
           date date_datetime year_month_firstDay
0           NaN           NaT                 NaT
1      2000-1-2    2000-01-02          2000-01-01
2      2000-1-3    2000-01-03          2000-01-01
3      2000-1-4    2000-01-04          2000-01-01
4      2000-1-5    2000-01-05          2000-01-01
..          ...           ...                 ...
223  2000-12-15    2000-12-15          2000-12-01
224  2000-12-16    2000-12-16          2000-12-01
225  2000-12-17    2000-12-17          2000-12-01
226  2000-12-18    2000-12-18          2000-12-01
227  2000-12-19    2000-12-19          2000-12-01

[228 rows x 3 columns]
>>> df.dtypes
date                           object
date_datetime          datetime64[ns]
year_month_firstDay    datetime64[ns]
dtype: object

Change datetime to the following month of the same day [2]

>>> df["year_nextMonth_firstDay"] = df["year_month_firstDay"] + pd.DateOffset(months=1)
>>> df
           date date_datetime year_month_firstDay year_nextMonth_firstDay
0      2000-1-1           NaT                 NaT                     NaT
1      2000-1-2    2000-01-02          2000-01-01              2000-02-01
2      2000-1-3    2000-01-03          2000-01-01              2000-02-01
3      2000-1-4    2000-01-04          2000-01-01              2000-02-01
4      2000-1-5    2000-01-05          2000-01-01              2000-02-01
..          ...           ...                 ...                     ...
223  2000-12-15    2000-12-15          2000-12-01              2001-01-01
224  2000-12-16    2000-12-16          2000-12-01              2001-01-01
225  2000-12-17    2000-12-17          2000-12-01              2001-01-01
226  2000-12-18    2000-12-18          2000-12-01              2001-01-01
227  2000-12-19    2000-12-19          2000-12-01              2001-01-01

[228 rows x 4 columns]
>>> df.dtypes
date                               object
date_datetime              datetime64[ns]
year_month_firstDay        datetime64[ns]
year_nextMonth_firstDay    datetime64[ns]
dtype: object

Pandas version used to build those test cases

>>> pd.__version__
'1.3.2'

Reference

[1] Extracting the first day of month of a datetime type column in pandas

[2] https://www.plus2net.com/python/pandas-dt-dateoffset.php