ENH: Allow dt accessor when using ArrowDtype with datetime types #50954
Conversation
|
Any other comments here @jbrockmendel ? |
|
will take another look |
| Callable to map the delegate's function to the cls' function. | ||
| raise_on_missing: bool, default True | ||
| Raise if an accessor does not exist on delegate. | ||
| False skips the missing accessor. |
There was a problem hiding this comment.
do you expect to reuse this machinery again? if not, could the pyarrow-dt be accessor be implemented standalone to avoid making this more complicaed?
There was a problem hiding this comment.
I am hoping to use this for the pyarrow duration types as well.
I think a standalone implementation would require a further refactoring of this dt accessor. Looks like currently datetime and timedelta dispatch to the Index while we probably need something that dispatches to the array
pandas/core/indexes/accessors.py
Outdated
| @@ -472,6 +559,8 @@ def __new__(cls, data: Series): | |||
| index=orig.index, | |||
| ) | |||
|
|
|||
| if isinstance(data.dtype, ArrowDtype) and data.dtype.kind in "mM": | |||
There was a problem hiding this comment.
Thanks yeah this should be "M" until pa.duration is supported
| fget=_getter, fset=_setter, doc=getattr(delegate, name).__doc__ | ||
| fget=_getter, | ||
| fset=_setter, | ||
| doc=getattr(delegate, accessor_mapping(name)).__doc__, |
There was a problem hiding this comment.
can we do this once near the top?
There was a problem hiding this comment.
ill put this into my next Assorted Cleanups branch
Enabled & tested for
pa.timestamptypes so far. Will follow up withpa.durationtypes in a subsequent PR