[
https://issues.apache.org/jira/browse/CALCITE-7463?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18074929#comment-18074929
]
Julian Hyde commented on CALCITE-7463:
--------------------------------------
My view is that the planner could run in an ARBITRARY_REPEATABLE mode where two
expressions that are arbitrary and structurally equal are considered to be
equivalent (i.e. always return the same results). In that mode, rewrites such
as the above would be valid.
If a rule needs ARBITRARY_REPEATABLE, it would have to declare it. The rule
would not fire in a planner where ARBITRARY_REPEATABLE is false.
We would need to classify expressions into deterministic, arbitrary, random.
(We might need to refine the distinction because an expression like 'select *
from emp order by deptno' has deterministic rows but arbitrary ordering;
'select * from emp order by deptno limit 3' has arbitrary rows and ordering.)
> UnionToFilterRule incorrectly rewrites UNION with LIMIT
> -------------------------------------------------------
>
> Key: CALCITE-7463
> URL: https://issues.apache.org/jira/browse/CALCITE-7463
> Project: Calcite
> Issue Type: Bug
> Components: core
> Affects Versions: 1.41.0
> Reporter: Zhen Chen
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.42.0
>
>
> The {{UnionToFilterRule}} produces incorrect results when applied to inputs
> that contain {{{}LIMIT{}}}.
> Specifically, the rule incorrectly collapses:
> {code:java}
> (SELECT mgr, comm FROM emp LIMIT 2)
> UNION
> (SELECT mgr, comm FROM emp LIMIT 2) {code}
> into:
> {code:java}
> SELECT DISTINCT mgr, comm FROM emp LIMIT 2 {code}
> This transformation is {*}not semantically equivalent{*}.
> *Reproduction*
> SQL
> {code:java}
> (SELECT mgr, comm FROM emp LIMIT 2)
> UNION
> (SELECT mgr, comm FROM emp LIMIT 2) {code}
> h4. Plan Before
> {code:java}
> LogicalUnion(all=[false])
> LogicalSort(fetch=[2])
> LogicalProject(MGR=[$3], COMM=[$6])
> LogicalTableScan(table=[[CATALOG, SALES, EMP]])
> LogicalSort(fetch=[2])
> LogicalProject(MGR=[$3], COMM=[$6])
> LogicalTableScan(table=[[CATALOG, SALES, EMP]]) {code}
> *Plan After (Incorrect)*
> {code:java}
> LogicalAggregate(group=[{0, 1}])
> LogicalSort(fetch=[2])
> LogicalProject(MGR=[$3], COMM=[$6])
> LogicalTableScan(table=[[CATALOG, SALES, EMP]]) {code}
> *Expected Behavior*
> The transformation should NOT be applied when any input of UNION contains
> LogicalSort(That contains ORDER BY, LIMIT, OFFSET).
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
