zhuqi-lucas commented on code in PR #21580:
URL: https://github.com/apache/datafusion/pull/21580#discussion_r3123905664
##########
datafusion/datasource-parquet/src/access_plan.rs:
##########
@@ -377,6 +381,124 @@ impl PreparedAccessPlan {
})
}
+ /// Return a reference to the row group indexes.
+ pub(crate) fn row_group_indexes(&self) -> &[usize] {
+ &self.row_group_indexes
+ }
+
+ /// Keep only the first `count` row groups, dropping the rest.
+ /// Used for TopK cumulative pruning after reorder + reverse.
+ pub(crate) fn truncate_row_groups(mut self, count: usize) -> Self {
Review Comment:
Good catch @xudong963, currently truncate_row_groups drops row_selection
entirely, which loses page-level pruning state for the retained RGs. I'll fix
this by skipping truncation when row_selection is present cumulative
prune will only apply when there's no page-level pruning active.
This is safe because page pruning is already reducing I/O within those RGs.
--
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]
