[
https://issues.apache.org/jira/browse/FLINK-27519?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17872773#comment-17872773
]
lincoln lee commented on FLINK-27519:
-------------------------------------
Some updates: after debugged with code, the nested over agg query works fine in
calcite.
Verified related case in
`RelOptRulesTest`:
{code}
@Test void testProjectToWindowRuleForNestedOver() {
HepProgramBuilder builder = new HepProgramBuilder();
builder.addRuleClass(ProjectToWindowRule.class);
HepPlanner hepPlanner = new HepPlanner(builder.build());
hepPlanner.addRule(CoreRules.PROJECT_TO_LOGICAL_PROJECT_AND_WINDOW);
final String sql = "select deptno, f1, f2 from (select *, last_value(deptno)
over (order by empno) f2\n"
+ "from (select *, first_value(deptno) over (order by empno) f1 from emp))\n";
sql(sql).withPlanner(hepPlanner)
.check();
}
LogicalProject(DEPTNO=[$7], F1=[$9], $2=[$10])
LogicalWindow(window#0=[window(order by [0] aggs [LAST_VALUE($7)])])
LogicalWindow(window#0=[window(order by [0] aggs [FIRST_VALUE($7)])])
LogicalTableScan(table=[[CATALOG, SALES, EMP]])
{code}
> Fix duplicates names when there are multiple levels of over window aggregate
> ----------------------------------------------------------------------------
>
> Key: FLINK-27519
> URL: https://issues.apache.org/jira/browse/FLINK-27519
> Project: Flink
> Issue Type: Bug
> Components: Table SQL / Planner
> Affects Versions: 1.15.0
> Reporter: Feng Jin
> Assignee: lincoln lee
> Priority: Major
> Fix For: 2.0.0
>
>
> A similar issue like
> [FLINK-22121|https://issues.apache.org/jira/browse/FLINK-22121]
> And can be reproduced by adding this unit test
> org.apache.flink.table.planner.plan.stream.sql.agg.GroupWindowTest#testWindowAggregateWithAnotherWindowAggregate
> {code:java}
> //代码占位符
> @Test
> def testWindowAggregateWithAnotherWindowAggregate(): Unit = {
> val sql =
> """
> |SELECT CAST(pv AS INT) AS pv, CAST(uv AS INT) AS uv FROM (
> | SELECT *, count(distinct(c)) over (partition by a order by b desc)
> AS uv
> | FROM (
> | SELECT *, count(*) over (partition by a, c order by b desc) AS pv
> | FROM MyTable
> | )
> |)
> |""".stripMargin
> util.verifyExecPlan(sql)
> } {code}
> The error message :
>
>
> {code:java}
> //代码占位符
> org.apache.flink.table.api.ValidationException: Field names must be unique.
> Found duplicates: [w0$o0] at
> org.apache.flink.table.types.logical.RowType.validateFields(RowType.java:273)
> at org.apache.flink.table.types.logical.RowType.<init>(RowType.java:158)
> at org.apache.flink.table.types.logical.RowType.of(RowType.java:298)
> at org.apache.flink.table.types.logical.RowType.of(RowType.java:290)
> at
> org.apache.flink.table.planner.calcite.FlinkTypeFactory$.toLogicalRowType(FlinkTypeFactory.scala:663)
> at
> org.apache.flink.table.planner.plan.nodes.physical.stream.StreamPhysicalOverAggregate.translateToExecNode(StreamPhysicalOverAggregate.scala:57)
> at
> org.apache.flink.table.planner.plan.nodes.exec.ExecNodeGraphGenerator.generate(ExecNodeGraphGenerator.java:74)
> at
> org.apache.flink.table.planner.plan.nodes.exec.ExecNodeGraphGenerator.generate(ExecNodeGraphGenerator.java:71)
> {code}
>
> I think we can add come logical in FlinkLogicalOverAggregate to avoid
> duplicate names of output rowType.
>
>
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
