JingsongLi commented on code in PR #208:
URL: https://github.com/apache/paimon-rust/pull/208#discussion_r3037566640
##########
crates/paimon/src/table/read_builder.rs:
##########
@@ -67,8 +68,11 @@ impl<'a> ReadBuilder<'a> {
/// Stats pruning is per file. Files with a different `schema_id`,
/// incompatible stats layout, or inconclusive stats are kept.
///
- /// [`TableRead`] does not evaluate row-level filters; callers must apply
- /// any remaining predicates themselves.
+ /// [`TableRead`] may use supported non-partition data predicates only on
+ /// the regular Parquet read path for conservative row-group pruning and
+ /// native Parquet row filtering. Unsupported predicates, non-Parquet
+ /// reads, and data-evolution reads remain residual and should still be
+ /// applied by the caller if exact filtering semantics are required.
pub fn with_filter(&mut self, filter: Predicate) -> &mut Self {
self.filter = Some(filter);
Review Comment:
Maybe here we can just split filter to partition_predicate and
data_predicates, and just pass them to TableScan and TableRead.
--
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]
