Skip to content

fix: bypass Iceberg tables in HiveTableValidator SD check (rebuilt glue-datacatalog-hive3-client jar) - #360

Merged
jamespfaulkner merged 2 commits into
mainfrom
fix/update-glue-datacatalog-hive3-client-jar
Jul 29, 2026
Merged

fix: bypass Iceberg tables in HiveTableValidator SD check (rebuilt glue-datacatalog-hive3-client jar) #360
jamespfaulkner merged 2 commits into
mainfrom
fix/update-glue-datacatalog-hive3-client-jar

Conversation

@jamespfaulkner

Copy link
Copy Markdown
Contributor

Summary

Rebuilds lib/aws-glue-datacatalog-hive3-client-3.4.0-WD-1.jar to include a defense-in-depth fix in HiveTableValidator.notApplicableTableType(): the required-StorageDescriptor-properties validator is now also bypassed when a table's table_type parameter is ICEBERG, not just when storage_handler is set.

This guards against Iceberg tables whose StorageDescriptor gets stripped by external processes (e.g. Glue's managed column-statistics service issuing an UpdateTable without InputFormat/OutputFormat/SerdeInfo), which previously caused get_table to fail with:

InvalidObjectException: StorageDescriptor#InputFormat cannot be null for table: <name>

No source changes in this repo — the jar is a prebuilt dependency checked into lib/. Verified via javap decompilation that the only functional bytecode change is the new isIcebergTable(Table) check in HiveTableValidator; all other class files differ only due to constant-pool churn from the full module recompile. Class file version is unchanged (Java 8 / major version 52).

Test plan

  • Confirm WD reads/writes succeed against an Iceberg table with table_type=ICEBERG and no storage_handler param, including one with a stripped StorageDescriptor
  • Confirm existing Hive/managed table validation behavior is unchanged
  • Run existing waggle-dance-core test suite

…tor bypass

Patches HiveTableValidator.notApplicableTableType() to also bypass required-SD
validation when table_type=ICEBERG, not just storage_handler. Iceberg tables
whose StorageDescriptor gets stripped (e.g. by the Glue managed column-stats
service) were failing get_table with "StorageDescriptor#InputFormat cannot be
null" instead of falling through the Iceberg bypass path.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
@jamespfaulkner jamespfaulkner changed the title fix: rebuild glue-datacatalog-hive3-client jar with Iceberg SD valida… fix: bypass Iceberg tables in HiveTableValidator SD check (rebuilt glue-datacatalog-hive3-client jar) Jul 29, 2026
@jamespfaulkner
jamespfaulkner marked this pull request as ready for review July 29, 2026 15:11
@jamespfaulkner
jamespfaulkner requested a review from a team as a code owner July 29, 2026 15:11
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
@jamespfaulkner
jamespfaulkner merged commit e87036d into main Jul 29, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants