malinjawi commented on code in PR #12131:
URL: https://github.com/apache/gluten/pull/12131#discussion_r3321079920
##########
gluten-substrait/src/main/resources/substrait/proto/substrait/algebra.proto:
##########
@@ -197,6 +197,18 @@ message ReadRel {
repeated DeleteFile delete_files = 3;
}
+ message DeltaReadOptions {
+ enum RowIndexFilterType {
+ KEEP_ALL = 0;
+ IF_CONTAINED = 1;
+ IF_NOT_CONTAINED = 2;
+ }
+ RowIndexFilterType row_index_filter_type = 1;
+ bool has_deletion_vector = 2;
+ uint64 deletion_vector_cardinality = 3;
+ uint32 deletion_vector_payload_index = 4;
+ }
Review Comment:
Done @zhztheplayer
I updated this PR to embed the serialized DV bytes directly in
`DeltaReadOptions` as `serialized_deletion_vector`.
This removes the separate payload index / side-band payload registry path.
The JVM now writes the materialized DV payload into the Substrait read options,
and Velox copies it into `DeltaSplitInfo`-owned storage before creating the DV
descriptor view.
I will keep deeper payload-size/perf work for the follow-up as you suggested
thanks.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
