Closed bryanculbertson closed 3 years ago
@kuanb I don't know this code, but the fix may be to change constraint_2
to use stop_times_by_stop
instead:
constraint_2 = (stop_times_by_stop.direction_id == direction)
direction_subset = stop_times_by_stop[constraint_2]
Same warnings occur for Pandas 1.1.5
Thanks, yeah I think the solution would be to actually access the boolean list as a numpy array rather than a pandas Series.
This would avoid even worrying about if we could swap out DataFrames (though I am 99 percent sure we can).
I just upgraded to Pandas
1.2.0
on Python3.8.5
and I am getting a warning from Pandas:At this line of code: https://github.com/kuanb/peartree/blob/254207ba759d82a1d66f84287075ce1f8d093463/peartree/parallel.py#L127
I think this is because the filter is using a different Dataframe from the Dataframe being filtered.