Open
Description
Feature Request / Improvement
Spark 4.0 introduces vectorized Parquet reading as the default behavior.
Currently the interfaces for Iceberg's Seekable Streams don't implement this behaviour.
This is fine because S3FileIO currently isn't setup to support Parquet's vectorized reader and the S3 client it makes use of doesn't perform well with parallel reads due to it not having coalescing, splitting, bounding logic . However the Analytics Accelerator for S3 does support this and can benefit from vectorized reads.
This issue is to add the necessary constructs and pass them through to AAL and the Parquet Reader.
Query engine
Spark
Willingness to contribute
- I can contribute this improvement/feature independently
- I would be willing to contribute this improvement/feature with guidance from the Iceberg community
- I cannot contribute this improvement/feature at this time