[
https://issues.apache.org/jira/browse/CALCITE-2840?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16767329#comment-16767329
]
Zoltan Haindrich commented on CALCITE-2840:
-------------------------------------------
Sorry for the unclear summary; I've rephrased it - in this case may was a bad
choice.
What I have in mind is that we are "carefully" setting UnknownAs to UNKNOWN
during and/or operand simplification; so that the following test is not passing
- because even if the expression is in UnknownAs.FALSE mode the "and" at the
bottom will be evaluating in UNKNOWN
{code}
@Test public void testSimplifyAnd4() {
checkSimplify2(
and(
eq(vInt(2), literal(2)),
or(
eq(vInt(3), literal(3)),
and(
ge(vInt(), literal(1)),
le(vInt(), literal(1))))),
"AND(=(?0.int2, 2), OR(=(?0.int3, 3), AND(>=(?0.int0, 1), <=(?0.int0,
1))))",
"AND(=(?0.int2, 2), OR(=(?0.int3, 3), =(?0.int0, 1)))"
);
}
{code}
There could be other cases; the important in this is that the dependent
simplification logic relies on the unknownAs mode.
> Simplification should use more specific UnknownAs modes during simplification
> -----------------------------------------------------------------------------
>
> Key: CALCITE-2840
> URL: https://issues.apache.org/jira/browse/CALCITE-2840
> Project: Calcite
> Issue Type: Improvement
> Reporter: Zoltan Haindrich
> Assignee: Zoltan Haindrich
> Priority: Major
>
> Some parts of this was already working in 1.17 ; but 1.18 turned out to be a
> more conservative in this sense.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
