I have a Pandas Dataframe with a date column. I want to only have the oldest records for each month and remove any records that came before. There will be duplicates and I want to keep them. I also need a new column with only the month and year.

Input

Output:

CodePudding user response：

Create column month_year with Series.dt.strftime and then compare maximal datetimes per groups by original date column in GroupBy.transform and filter in boolean indexing:

df['date'] = pd.to_datetime(df['date'], dayfirst=True)

df = df.assign(month_year = df['date'].dt.strftime('%m/%Y'))

df = df[df.groupby(['Provider', 'month_year'])['date'].transform('max').eq(df['date'])]
print (df)
  Provider       date month_year
2    Apple 2022-01-20    01/2022
3    Apple 2022-01-20    01/2022
5    Apple 2022-02-10    02/2022