xarray
xarray copied to clipboard
pydata
Harmonize returned multi-indexed indexes when applying `concat` along new dimension
- [x] Closes #6881
- [x] Tests added
In the current implementation, the concat function does not ensure that the indexes that belong to the same MultiIndex relate to the same MultiIndex object in Dataset.indexes when concatenating along a new dimension. This becomes a problem as soon as the returned dataset needs to be aligned (for broadcasting, reindexing etc.), see #6881 for example. As far as I understand, this bug was introduced in https://github.com/pydata/xarray/pull/5692 following the idea that the concat function should disentangle indexes and dimensions.
It can be fixed by not removing the index name from the list of indexes which should be merged, see https://github.com/pydata/xarray/blob/9050a8b9efc28142b762475c7285603a87b00e83/xarray/core/concat.py#L493. All indexes contained in this list will get a new index object. Currently, this list only contains the levels of a multi-indexed index, not the index name itself. This is removed as it is contained in dim_names. Instead of removing all dimension names from this list, I suggest to only remove unlabeled dimensions from this list.
@benbovy I applied your suggestion (+ a small fix) and merged with the up to date main branch. All tests pass.
LGTM, thanks @FabianHofmann! Could you update whats-new.rst in the docs? I think this bug fix is worth to be mentioned there with your contribution.
