ramitg254 commented on code in PR #6413:
URL: https://github.com/apache/hive/pull/6413#discussion_r3378741694
##########
ql/src/java/org/apache/hadoop/hive/ql/metadata/Table.java:
##########
@@ -595,26 +606,46 @@ public boolean equals(Object obj) {
&& Objects.equals(snapshotRef, other.snapshotRef);
}
+
+ /**
+ * Returns partition columns, consulting the storage handler for non-native
tables (e.g. Iceberg)
+ * where partition columns are not stored in the metastore.
+ */
public List<FieldSchema> getPartCols() {
- List<FieldSchema> partKeys = tTable.getPartitionKeys();
- if (partKeys == null) {
- partKeys = new ArrayList<>();
- tTable.setPartitionKeys(partKeys);
+ if (tablePartCols != null) {
+ return tablePartCols;
+ }
+ if (isTableTypeSet() && hasNonNativePartitionSupport()) {
Review Comment:
this check is guard to find out is iceberg table registered yet even though
its Storage Handler type is defined as iceberg.
without this guard whenever in operations like alter table convert to
iceberg, some checks are done before registering the new iceberg table during
execution like one of them includes is `checkValidity(table, context);`
which eventually calls `getPartCols()` and since iceberg table is not
registered yet so `table type prop is not set to iceberg` and it will raise
`NoSuchIcebergTableException` when called
`getStorageHandler().getPartitionKeys(this)`
so to avoid these calls for any kind of iceberg tables which are not yet
registered this guard is added
--
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]
