Inconsistent treatment of id in roll_time_series

[david-waterworth]

I'm not sure if this is a bug, but the internal helper in roll_time_series hard codes the id colume regardless of the value of the input argument column_id

tsfresh/tsfresh/utilities/dataframe_functions.py

Line 346 in da06274

df_temp["id"] = df_temp[column_id].apply(lambda row: (row, timeshift_value))

This means that it either replaces the original id (column_id=='id'), or retains it (column_id!='id') which seems sligtly inconsistent?

[nils-braun]

That is correct @david-waterworth - thanks for spotting this.
I guess it should be replaced by column_id - keeping the original name makes the most sense to me. Would you like to do a PR for this?

[david-waterworth]

I started on a PR and noticed another hard-coded references, i.e. in roll_time_series the final results are sorted by "id" i.e.

return df_shift.sort_values(by=["id", column_sort or "sort"])

I'm wondering if this should be

return df_shift.sort_values(by=[column_id or "id", column_sort or "sort"])

[nils-braun]

If you are using column_id in the helper function, the result should only have column_id and not id anymore or did I miss something? Feel free to do the PR anyways and we can discuss more easily in the review alongside the code :-)

[nils-braun]

Ah sorry, you did already create a PR (I went through my notifications from top to bottom...) :-)