[GitHub] [arrow] lhoestq commented on pull request #33925: GH-33923: [Docs] Tensor canonical extension type specification

["lhoestq (via GitHub)" <gi...@apache.org>]

lhoestq commented on PR #33925:
URL: https://github.com/apache/arrow/pull/33925#issuecomment-1416115543

   Regarding what `dim_names` would imply for `pyarrow` and `numpy`:
   
   If I have a numpy array, convert it to arrow using `pa.array`, and then call `to_numpy()`, ideally I would expect to get the exact same numpy array.
   
   If we use `dim_names` and the array has a physical order that is different that the logical order (but still row-major), does that mean I would have to give names to the dimensions to do [numpy -> arrow] and to do [arrow -> numpy] to end up with the same array back ?
   
   Arrow could also give arbitrary names and save the original logical order for the dimensions.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org