February 14, 2023
This release of the Cloudera Data Warehouse (CDW) service on CDP Public Cloud has the following known issues:
Technical Service Bulletins
- TSB-732 2024: Incorrect results are generated by Hive JOIN when bloom filter is activated
- The bloom filter implemented in HIVE-23880 was designed to enhance performance for queries with JOIN statements, where one small table and another significantly larger table is joined on partition keys. However, the bloom filter introduced an issue in Apache Hive (Hive), when dynamic semijoin redaction is involved that generates incorrect query results. This issue is corrected in HIVE-26655.
- Upstream JIRA
- Hive-23880(cause)HIVE-26655(fix)
- Knowledge article
- For the latest update on this issue see the corresponding Knowledge article: TSB 2024-732: Incorrect results are generated by Hive JOIN when bloom filter is activated
