Feature Request / Improvement
We are creating tables with large size columns (up to hundreds Mbs), these large size columns normally are some bytes where stats (min/max) does not make sense. Today, for Iceberg tables' data files in parquet, columns stats is enabled by default. The stats for these large size columns consumes lots of memory. More background information is at parquet PR.
apache/parquet-java#2521
Similar to metrics, we need table properties to enable/disable parquet column statistics.
Query engine
None
Willingness to contribute
Feature Request / Improvement
We are creating tables with large size columns (up to hundreds Mbs), these large size columns normally are some bytes where stats (min/max) does not make sense. Today, for Iceberg tables' data files in parquet, columns stats is enabled by default. The stats for these large size columns consumes lots of memory. More background information is at parquet PR.
apache/parquet-java#2521
Similar to metrics, we need table properties to enable/disable parquet column statistics.
Query engine
None
Willingness to contribute