[Caideyipi]

Description

IoTDBTimeDurationIT.testTimeDurationManyColumnOnManyDataRegion compares GROUP BY TAGS rows in lockstep across DataNodes, although the query does not guarantee group order. Valid results with different Beijing/Nanjing row orders can fail in ClusterTestResultSet.getString() before the test's expected-result assertion runs.

Check this query through a connection pinned to each DataNode and match the expected rows independently of their order. Preserve checks for column headers, aggregation values, and missing, duplicate, or unexpected rows.

Validation

• mvn -o spotless:apply -pl integration-test -P with-integration-tests passed.
• Integration-test main and test sources compiled successfully; Checkstyle and Spotless checks passed.
• IoTDBTimeDurationIT#testTimeDurationManyColumnOnManyDataRegion passed with ClusterIT and HighPerformanceMode (1 ConfigNode, 3 DataNodes): 1 test, 0 failures, 0 errors. The complete module verify lifecycle passed, using locally installed 2.0.11-SNAPSHOT runtime dependencies.

This PR has:

• been self-reviewed.
• modified the affected integration test.
• been tested in a test IoTDB cluster.