Closed
Description
Once #9372 goes in we will support grouping by multiple variables.
@max-sixty makes a good point re: ergonomics:
I missed previous discussions / can decide this later — but do we want to have
da.groupby(['foo', 'bar'])
as sugar forda.groupby(foo=UniqueGrouper(), bar=UniqueGrouper())
?
Originally posted by @max-sixty in #9372 (comment)
I think this would be great. da.groupby(foo=UniqueGrouper(), bar=UniqueGrouper())
is a bit too much typing for a very simple expression.