group by rank continuous month by pandas

Question

This is something different by this postenter link description here.

Example

ID    TIME
01    2018-07-01
01    2018-08-01
01    2018-09-01         


        

Answer 1

You can convert datetimes to months periods by Series.dt.to_period and then compare by MonthEnd difference of values by DataFrameGroupBy.diff with cumulative sums and last is used GroupBy.cumcount:

df['TIME'] = pd.to_datetime(df['TIME']).dt.to_period('M')
new = df.groupby('ID', group_keys=False)['TIME'].diff().ne(pd.offsets.MonthEnd()).cumsum()
df['rank'] = df.groupby(['ID',new]).cumcount().add(1)
print (df)
  ID     TIME  RANK  rank
0   1  2018-07     1     1
1   1  2018-08     2     2
2   1  2018-09     3     3
3   1  2018-11     1     1
4   1  2018-12     2     2
5   1  2019-01     3     3
6   2  2019-01     1     1
7   2  2019-02     2     2
8   2  2019-12     1     1
9   2  2020-01     2     2