Known limitations

While we consider xarray-events to be useful for some specific use cases, we do also recognize that it still has a lot of room for improvement. This library is meant to be applicable to a wide range of areas, so we have attempted to maintain generality in mind while developing it all along.

In this section we detail some known limitations of this library.

Custom-type duration values

Tracking bug: https://github.com/teibit/xarray-events/issues/2

The values of the events DataFrame on the columns that specify a duration should be integers. This means that they must not be custom objects. This is worth mentioning because one might want them to be Timestamp objects (from Pandas), for example, but this is not possible because comparison between them isn’t well implemented. This limitation isn’t impactful until one starts using groupby_events() because of df_contains_overlapping_events().