924060929 commented on code in PR #67182:
URL: https://github.com/apache/doris/pull/67182#discussion_r3919940106
##########
fe/fe-connector/fe-connector-hive/src/main/java/org/apache/doris/connector/hive/HiveConnectorMetadata.java:
##########
@@ -577,8 +577,9 @@ public ConnectorTableSchema getTableSchema(
if (supportsHiveSampleAnalyze(tableInfo)) {
perTableCapabilities.add(ConnectorCapability.SUPPORTS_SAMPLE_ANALYZE);
}
- if (supportsHiveTopNLazyMaterialize(tableInfo)) {
+ if (supportsHiveOrcOrParquetScan(tableInfo)) {
perTableCapabilities.add(ConnectorCapability.SUPPORTS_TOPN_LAZY_MATERIALIZE);
+
perTableCapabilities.add(ConnectorCapability.SUPPORTS_STORAGE_PREDICATE_PRUNING);
Review Comment:
Thanks for pointing this out. The route analysis is valid:
`HoodieParquetInputFormatBase` is treated as plain Hive and uses the native
Parquet reader, while the current exact-class gate does not grant it
`SUPPORTS_STORAGE_PREDICATE_PRUNING`.
This is an optimization-coverage gap rather than a correctness issue. The
original function predicate is retained, so query results are unchanged; this
route only misses the newly inferred bare-column bounds and may scan more
Parquet row groups. It also does not regress the behavior that existed before
this PR.
We accept this limited P2 gap for the current PR and will keep the current
change unchanged; the special Hoodie-base admission can be handled separately
from the Top-N whitelist.
--
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]
