[IMPALA-7755] IS [NOT] DISTINCT FROM rewrite rules don't handle aggregates

• Type: Bug
• Status: Open
• Priority: Minor
• Resolution: Unresolved
• Affects Version/s: Impala 3.0
• Component/s: Frontend
• Epic Color: ghx-label-9
• Assignee: Unassigned
• Reporter: Paul Rogers
• Created: 25/Oct/18 03:27
• Updated: 29/Oct/18 20:01

Description

Testing revealed one issue and one limitation of the IS [NOT] DISTINCT FROM optimizations in SimplifyDistinctFromRule.

The rule can simplify sum(id) <=> sum(id) to false. But, according to IMPALA-5125, the rewrite cannot be applied if it would remove the only aggregate. This is a bug.

A simplification is missing: id IS DISTINCT FROM NULL → id IS NOT NULL.

Here is a specific test case from FullRewriteTest create as part of IMPALA-7655:

  @Test
  public void TestSimplifyDistinctFromRule() throws ImpalaException {
    verifySelectRewrite("if(sum(int_col) <=> sum(int_col), 1, 2)",
        "CASE WHEN sum(int_col) IS NOT DISTINCT FROM sum(int_col) THEN 1 ELSE 2 END");
  }

The above fails, the re-writer produces "2" because it ignores the "must preserve aggregate" rule.

Issue Links

• is part of: IMPALA-7747 Clean up the Expression Rewriter