My purpose is to group the datetime indexed dataframe into months and write to a parquet file but chunked / row_grouped by month.

The reason for this is that I have many dataframes, and each dataframe will be a parquet file. However, when I read those dataframes, I don’t need to always read the whole, yet just read several months of each.

For example, one typical dataframe I have is like

                    Level  Score
2023-10-21 00:01:00     A      1
2023-10-22 00:01:00     B      2
...
2023-11-01 00:01:00     C      3
...
2023-11-30 00:01:00     D      4
2023-12-01 00:02:00     A      5
...
2023-12-22 00:02:00     B      6

I would like have one parquet file for this dataframe, yet, inside somehow group / chunk the parquet by month, so I can read, e.g., only 2023-10 data from this file without touching other months.

I am reading the pyarrow parquet. I see parquet has chunksize, row groups etc.

However, I got confused by them.

How can I group the chunks / rows by month?