[
https://issues.apache.org/jira/browse/CALCITE-5939?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17756426#comment-17756426
]
LakeShen commented on CALCITE-5939:
-----------------------------------
Hi [~suibianwanwan33] ,your idea is very good.
Could you give an SQL example of a scenario where materialized view rewriting
is supported by SR, but not by Calcite.
> Support join derivability mv rewrite for unifyRule
> --------------------------------------------------
>
> Key: CALCITE-5939
> URL: https://issues.apache.org/jira/browse/CALCITE-5939
> Project: Calcite
> Issue Type: New Feature
> Components: core
> Reporter: kate
> Priority: Minor
>
> Currently more and more materialized view matching algorithms have been able
> to support the matching of different join types, can we try to support some
> optimization based on the existing UnifyRule to do
> Here are the materialized matching capabilities of different join types in
> StarRocks
> |mv|query|condition|
> |LEFT OUTER JOIN ON A.a1 = B.b1|INNER JOIN ON A.a1 = B.b1|NULL|
> |RIGHT OUTER JOIN ON A.a1 = B.b1|INNER JOIN ON A.a1 = B.b1|NULL|
> |LEFT OUTER JOIN ON A.a1 = B.b1|LEFT ANTI JOIN ON A.a1 = B.b1|NULL|
> |RIGHT OUTER JOIN ON A.a1 = B.b1|RIGHT ANTI JOIN ON A.a1 = B.b1|NULL|
> |INNER JOIN ON A.a1 = B.b1|LEFT SEMI JOIN ON A.a1 = B.b1|The right table must
> have a unique key or a primary key |
> |INNER JOIN ON A.a1 = B.b1|RIGHT ANTI JOIN ON A.a1 = B.b1|The left table must
> have a unique key or a primary key|
> |FULL OUTER JOIN ON A.a1 = B.b1|LEFT OUTER JOIN ON A.a1 = B.b1|At least one
> column of the left table is notnull|
> |FULL OUTER JOIN ON A.a1 = B.b1|NNER OUTER JOIN ON A.a1 = B.b1|At least one
> column of the right table is not null|
> |FULL OUTER JOIN ON A.a1 = B.b1|INNER OUTER JOIN ON A.a1 = B.b1|At least one
> column of the left and right table is notnull|
>
> Is it possible to enhance calcite's ability to rewrite materialization
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
