[
https://issues.apache.org/jira/browse/CALCITE-4800?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17484056#comment-17484056
]
Xurenhe commented on CALCITE-4800:
----------------------------------
Handle for Sort's mv-match
* Equal collation's info, we just check offset and fetch. Query could be
rewritten by mv, when mv's data-range must be larger than query's, such as:
{code:java}
--query:
Sort(collation: [1], offset: 3)
Calc(program: (expr#0..4=[{inputs}], name=[$t2], deptno=[$t1]))
Scan(table: [hr, emps])
--mv:
Sort(collation: [1], offset: 2)
Calc(program: (expr#0..4=[{inputs}], name=[$t2], deptno=[$t1]))
Scan(table: [hr, emps])
--rewrite:
Sort(collation: [], offset: 1)
Sort(collation: [1], offset: 2)
Calc(program: (expr#0..4=[{inputs}], name=[$t2], deptno=[$t1]))
Scan(table: [hr, emps])
--replace:
Sort(collation: [], offset: 1)
Scan(table: [mv]) {code}
* No-equal collation's info, we could build a new sort above mv's, which
collation must be query', such as:
{code:java}
--query:
Sort(collation: [1 DESC])
Calc(program: (expr#0..4=[{inputs}], name=[$t2], deptno=[$t1]))
Scan(table: [hr, emps])
--mv:
Sort(collation: [1])
Calc(program: (expr#0..4=[{inputs}], name=[$t2], deptno=[$t1]))
Scan(table: [hr, emps])
--rewrite:
Sort(collation: [1 DESC])
Sort(collation: [1])
Calc(program: (expr#0..4=[{inputs}], name=[$t2], deptno=[$t1]))
Scan(table: [hr, emps])
--replace:
Sort(collation: [1 DESC])
Scan(table: [mv]) {code}
PR: [https://github.com/apache/calcite/pull/2702]
[~xzh_dz] Hi please review this pr, if you have time.
Thanks
> Add SortToSortUnifyRule in SubstitutionVisitor
> ----------------------------------------------
>
> Key: CALCITE-4800
> URL: https://issues.apache.org/jira/browse/CALCITE-4800
> Project: Calcite
> Issue Type: Improvement
> Reporter: xzh_dz
> Assignee: Xurenhe
> Priority: Major
> Labels: pull-request-available
> Time Spent: 40m
> Remaining Estimate: 0h
>
> In my project, there are more usage scenarios for materialized view
> recognition, and the `Sort Operator` can help sort data sets.But it has not
> been supported yet.A `Sort Operator` failed example can be reproduced as
> follows.
> {code:java}
> // code placeholder
> org.apache.calcite.test.MaterializedViewSubstitutionVisitorTest
> @Test void testSortToSort() {
> final String mv = ""
> + "select \"name\", \"deptno\"\n"
> + "from \"emps\"\n"
> + "order by \"deptno\"";
> final String query = ""
> + "select \"name\", \"deptno\"\n"
> + "from \"emps\"\n"
> + "order by \"deptno\" desc";
> sql(mv, query).withChecker(
> resultContains(""
> + "EnumerableTableScan(table=[[hr, MV0]])")).ok();
> }
> {code}
> Exception:
> {code:java}
> // code placeholder
> Materialized view failed to be matched by optimized results:Materialized view
> failed to be matched by optimized results:
> java.lang.AssertionError: Materialized view failed to be matched by optimized
> results:
> at
> org.apache.calcite.test.AbstractMaterializedViewTest.checkMaterialize(AbstractMaterializedViewTest.java:116)
> at
> org.apache.calcite.test.AbstractMaterializedViewTest.access$000(AbstractMaterializedViewTest.java:67)
> at
> org.apache.calcite.test.AbstractMaterializedViewTest$Sql.ok(AbstractMaterializedViewTest.java:229)
> at java.lang.invoke.MethodHandle.invokeWithArguments(MethodHandle.java:627)
> at
> org.apache.calcite.util.ImmutableBeans.lambda$makeDef$3(ImmutableBeans.java:299)
> {code}
>
>
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
