geoffreyclaude commented on code in PR #15563:
URL: https://github.com/apache/datafusion/pull/15563#discussion_r2033541557
##########
datafusion/physical-expr/src/equivalence/properties/mod.rs:
##########
@@ -575,10 +576,35 @@ impl EquivalenceProperties {
// From the analysis above, we know that `[a ASC]` is satisfied.
Then,
// we add column `a` as constant to the algorithm state. This
enables us
// to deduce that `(b + c) ASC` is satisfied, given `a` is
constant.
- eq_properties = eq_properties
-
.with_constants(std::iter::once(ConstExpr::from(normalized_req.expr)));
+ eq_properties = eq_properties.with_constants(std::iter::once(
+ ConstExpr::from(Arc::clone(&normalized_req.expr)),
+ ));
}
- true
+
+ // All requirements are satisfied.
+ normalized_reqs.len()
+ }
+
+ /// Determines the longest prefix of `reqs` that is satisfied by the
existing ordering.
+ /// Returns that prefix as a new `LexRequirement`, and a boolean
indicating if all the requirements are satisfied.
+ pub fn extract_matching_prefix(
+ &self,
+ reqs: &LexRequirement,
+ ) -> (LexRequirement, bool) {
Review Comment:
Are you suggesting recomputing the normalized prefix for each pass inside
`attempt_early_completion`? Or using the un-normalized prefix instead (which
might result in longer bytes arrays to compare, if there were redundant columns
in the sort expression)?
I totally agree that storing only the count of common sort expressions would
be cleaner, but I don't see how to do that without additional compute cost...
--
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: github-unsubscr...@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org
For additional commands, e-mail: github-h...@datafusion.apache.org
