neilconway commented on code in PR #21362: URL: https://github.com/apache/datafusion/pull/21362#discussion_r3066706657
########## datafusion/sqllogictest/test_files/order.slt: ########## @@ -1521,7 +1579,7 @@ query TT EXPLAIN SELECT c1, c2 FROM table_with_ordered_pk ORDER BY c1, c2; ---- logical_plan -01)Sort: table_with_ordered_pk.c1 ASC NULLS LAST, table_with_ordered_pk.c2 ASC NULLS LAST +01)Sort: table_with_ordered_pk.c1 ASC NULLS LAST Review Comment: Hmm, the optimization seems sound to me. If c1 functionally determines c2, we know that each distinct c1 value is associated with exactly one c2 value. So sorting by c1 is sufficient; adding in c2 as a tiebreaker / secondary sort key is never useful. In the example, if `c1` is a PK of `table`, sorting by c1 is sufficient to get the right sort order -- there will never be two rows with the same `c1` value. -- 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]
