mchades commented on code in PR #13771:
URL: https://github.com/apache/gravitino/pull/13771#discussion_r4248005625
##########
core/src/main/java/org/apache/gravitino/catalog/TableOperationDispatcher.java:
##########
@@ -582,11 +592,12 @@ private EntityCombinedTable importTable(NameIdentifier
identifier) {
.withAuditInfo(audit)
.build();
try {
- // Overwrite only with the id stored in the catalog: it identifies the
table, so a row under
- // the same name with another id is stale and gets replaced. A generated
id identifies
- // nothing, so a row that appeared meanwhile, e.g. from a concurrent
import on another node,
- // must win. The plain insert then conflicts, and loadTable reloads that
row.
- store.put(tableEntity, stringId != null);
+ // Never overwrite imported table metadata. In particular, an overwrite
by stable ID can
+ // silently change schema or catalog ownership and make existing
privileges refer to the
+ // wrong path.
+ store.put(tableEntity, false);
Review Comment:
**[P1] Normalize Hive-backed Iceberg aliases before rejecting imports**
With the Hive backend of `lakehouse-iceberg`, loading a table through an
uppercase spelling after creating it with a lowercase name resolves to the same
Hive table and `gravitino.identifier`. `IcebergCatalogCapability` still uses
the default case-sensitive rules, so both the initial lookup and the conflict
reload use the uppercase path. This create-only write then conflicts with the
existing ID, and loading the table fails with a manual-reconciliation error.
The `INSERT INTO %S` in `SparkCommonIT.testAlterTableReplaceColumns` already
exercises this scenario: the [historical CI
log](https://github.com/apache/gravitino/actions/runs/38220570510/job/114730549951)
shows the failure while loading `TEST_REPLACE_COLUMNS_TABLE`. The new
`SparkIcebergCatalogHiveBackendIT` override changes that normal write to an
expected failure. Normalize identifiers according to the Hive backend before
rejecting the import, and retain the original success assertions so ordinary
mixed-case SQL continues to work.
--
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]
