[BUG] Raise error when Ray Data tensor cannot be pickled and disable compliant nested types #2428
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This solves some of our flaky unit tests that try to read a Ray dataset tensor type under a Pyarrow version that does not properly serialize it. Adds a new exception when we detect this issue, and also updates our pyarrow dependencies to 13.0.0.
I also found a separate bug upon upgrading to pyarrow, where versions >= 13.0.0 will use compliant Parquet nested type names by default (see apache/arrow#35146), which would then cause a discrepancy between the Parquet and Arrow schemas for extension types. This would lead to incorrect conversion from Parquet to Arrow when we would read a file. This PR disables that explicitly.
Finally I also added a
coerce_temporal_nanoseconds
parameter toto_pandas
to revert it to its pre pyarrow>=13.0.0 behavior, which makes the sql integration tests pass againConfirmed to raise proper error in local testing