Details
-
Improvement
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
Description
In ARROW-11843 we found that we still can't enable parallel column conversion for Parquet datasets because the Arrow Parquet reader uses OptionalParallelFor, which does a parallel wait. We should provide an async version to avoid nested parallelism.
Attachments
Issue Links
- is duplicated by
-
ARROW-12916 [C++][Parquet] Always transfer to executor in Parquet AsyncGenerator
- Closed
- relates to
-
ARROW-11843 [C++] Add asynchronous read to parquet::arrow::FileReader
- Resolved
- links to