[SPARK-53840][SQL][FollowUp] Make SHOW TABLE EXTENDED AS JSON match V2 metadata

[linhongliu-db]

What changes were proposed in this pull request?

This follow-up updates SHOW TABLE EXTENDED ... AS JSON for V2 catalogs to use catalog-provided summary metadata when producing extended JSON rows.

• For V2 relation catalogs, extended JSON now uses RelationCatalog#listRelationSummaries, so relations such as views are preserved.
• For V2 table catalogs, extended JSON now uses TableCatalog#listTableSummaries.
• The extended JSON type field now comes from TableSummary.tableType() instead of being hard-coded to TABLE.
• V2 command-suite coverage now checks the V2 summary table type and covers a relation catalog returning both a table and a view.

Why are the changes needed?

SHOW TABLE EXTENDED text output uses V2 summary metadata for table type information, but the JSON path was falling back to identifier-only table listing and hard-coding the type as TABLE. That loses V2 table type metadata, and for relation catalogs it can also miss V2 views that are available through relation summaries.

Does this PR introduce any user-facing change?

Yes. On unreleased Spark master, SHOW TABLE EXTENDED ... AS JSON for V2 catalogs now reports the catalog-provided table or view type metadata and preserves V2 relation-catalog views in the JSON output.

How was this patch tested?

Added regression coverage in ShowTablesSuite.

Also checked:

git diff --check apache/master..HEAD

GitHub PR checks passed on head 39f507910201d5dd08bbdb58125e1f7c0ac3809a:

• Build
• Notify test workflow

Attempted local Spark SQL suite validation, but the environment could not download required build dependencies:

./build/sbt "sql/testOnly org.apache.spark.sql.execution.command.v2.ShowTablesSuite -- -z 'show table extended as json returns relation catalog table and view types'"
./build/mvn -pl sql/core -am -DskipTests -DskipJavaTests -DskipScalaTests -DskipMima -DskipCheckstyle test-compile
MAVEN_MIRROR_URL=https://repo.maven.apache.org/maven2 ./build/mvn -pl sql/core -am -DskipTests -DskipJavaTests -DskipScalaTests -DskipMima -DskipCheckstyle test-compile

Was this patch authored or co-authored using generative AI tooling?

Generated-by: Codex (GPT-5)

[uros-b]

LGTM, thank you @linhongliu-db! cc @cloud-fan @urosstan-db

[cloud-fan]

0 blocking, 0 non-blocking, 0 nits.
The implementation and regression coverage are consistent with the V2 catalog contracts; no actionable findings.

Verification

Traced both V2 catalog branches through listRelationSummaries / listTableSummaries, verified pattern filtering remains on relation names, compared the relation path with ShowTablesExec, and checked the TableCatalog and RelationCatalog default summary contracts. No tests were run as part of this review.

[cloud-fan]

Merge Summary:

• merged into master e6df455
• merged into branch-4.x 3818e53

Posted by merge_spark_pr.py