github-actions[bot] commented on code in PR #62315:
URL: https://github.com/apache/doris/pull/62315#discussion_r3062260788
##########
regression-test/suites/nereids_rules_p0/column_pruning/string_length_column_pruning.groovy:
##########
@@ -299,6 +299,21 @@ suite("string_length_column_pruning") {
notContains "type=bigint"
}
+ // ─── Struct field empty-string rewrite cases
────────────────────────────────
+
+ // element_at(struct_col, 'f3') = '' rewrites to
length(element_at(struct_col, 'f3')) = 0
+ // → OFFSET optimization applies to the struct string field
+ explain {
+ sql "select 1 from slcp_str_tbl where element_at(struct_col, 'f3') =
''"
+ contains "length"
Review Comment:
This only proves that the predicate was rewritten to `length(...)`. It does
not prove the behavior this PR is trying to unlock: the OFFSET-only
nested-column read for the struct string field. If nested-column pruning
regressed and still scanned the full field data, this test would still pass.
Please assert the actual access-path shape here too (`nested columns` +
`OFFSET`), like the other optimization checks in this suite.
--
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]
