public class HiveInsertOnlyScanWriteIdRule
extends org.apache.calcite.plan.RelOptRule
This rule turns on populating writeId of insert only table scans.
Currently fetching writeId from insert-only tables is not turned on automatically:
1. only not compacted records has valid writeId.
2. the writeId and bucketId is populated into the ROW_ID struct however the third field
of the struct (rowId) is always 0.
This feature is only used when rebuilding materialized view incrementally when the view has
insert-only source tables.