[trino-hudi connector] Support skipping based on record index

[hudi-bot]

JIRA info

• Link: https://issues.apache.org/jira/browse/HUDI-9296
• Type: Task
• Epic: https://issues.apache.org/jira/browse/HUDI-4394
• Fix version(s):
  • 1.1.0

[nsivabalan]

This appears to be implemented on current master. The Trino-Hudi connector now supports record-level-index based file-slice skipping via HudiRecordLevelIndexSupport (see hudi-trino-plugin/src/main/java/io/trino/plugin/hudi/query/index/HudiRecordLevelIndexSupport.java), wired into IndexSupportFactory and toggleable via the record_level_index_enabled session property. The strategy reads the record-index partition of the metadata table (HoodieTableMetadata.readRecordIndex) for the equality/IN predicates on record-key columns and prunes file slices by fileId. Landed in commit 2669933 (HUDI-9767, PR #13787). If this matches the original ask, can we close this issue / mark HUDI-9296 as Done? If something specific is still missing, happy to dig in further.