[HIVE-23038] Suspect Direct SQL Statement Regarding BitVector

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 3.1.2, 4.0.0
• Component/s: Standalone Metastore, Statistics
• Assignee: Unassigned
• Reporter: David Mollitor
• Created: 18/Mar/20 01:35
• Updated: 18/Mar/20 01:35

Description

https://github.com/apache/hive/blob/26cc3154c061d2194fba1c3bb156bb7e06e4a6c5/standalone-metastore/metastore-server/src/main/java/org/apache/hadoop/hive/metastore/MetaStoreDirectSql.java#L1249

The list of things that gets SELECT-ed from the backend database depends on a flag:

    final String queryText0 = "select " + getStatsList(enableBitVector) + " from " + TAB_COL_STATS
          + " where \"CAT_NAME\" = ? and \"DB_NAME\" = ? and \"TABLE_NAME\" = ? "
          + " and \"ENGINE\" = ? and \"COLUMN_NAME\" in (";

However, the same flag is not passed to the Java Marshaling code, so I sincerely doubt that it is being parsed correctly. How can it know what the number of columns are?