You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, if you have "a" pyarrow array object that can be either array or chunked array, and you call to_numpy with something that cannot be zero copy, you need to do:
because the chunked array version doesn't accept the keyword. To make this easier, I think we should just accept that keyword in ChunkedArray.to_numpy() as well (and only accept False as a possible value)
Component(s)
Python
The text was updated successfully, but these errors were encountered:
…mpy (#35582)
### Rationale for this change
Addresses GH-34787.
### What changes are included in this PR?
I think this proposes the minimal viable changes to have the same signature as proposed in GH-34787.
### Are these changes tested?
Yes, for now, a single simple test has been introduced to test the correct behavior of `ChunkedArray.to_numpy` with respect to the addition of `zero_copy_only`.
* Closes: #34787
Lead-authored-by: Julien Jerphanion <git@jjerphan.xyz>
Co-authored-by: Joris Van den Bossche <jorisvandenbossche@gmail.com>
Signed-off-by: Joris Van den Bossche <jorisvandenbossche@gmail.com>
Describe the enhancement requested
Currently, if you have "a" pyarrow array object that can be either array or chunked array, and you call
to_numpy
with something that cannot be zero copy, you need to do:instead of simply:
because the chunked array version doesn't accept the keyword. To make this easier, I think we should just accept that keyword in
ChunkedArray.to_numpy()
as well (and only accept False as a possible value)Component(s)
Python
The text was updated successfully, but these errors were encountered: