snuyanzin commented on code in PR #26840:
URL: https://github.com/apache/flink/pull/26840#discussion_r2239909439
##########
flink-table/flink-table-planner/src/test/resources/org/apache/flink/table/planner/plan/stream/sql/agg/OverAggregateTest.xml:
##########
@@ -352,50 +352,50 @@ Calc(select=[c, w0$o0 AS $1])
]]>
</Resource>
</TestCase>
- <TestCase name="testRowTimeBoundedPartitionedRangeOver">
+ <TestCase name="testRowTimeBoundedNonPartitionedRowsOver">
<Resource name="sql">
<![CDATA[
SELECT c,
- COUNT(a) OVER (PARTITION BY c ORDER BY rowtime
- RANGE BETWEEN INTERVAL '1' SECOND PRECEDING AND CURRENT ROW)
- FROM MyTable
+ COUNT(a) OVER (ORDER BY rowtime ROWS BETWEEN 5 PRECEDING AND CURRENT ROW)
+FROM MyTable
]]>
</Resource>
<Resource name="ast">
<![CDATA[
-LogicalProject(c=[$2], EXPR$1=[COUNT($0) OVER (PARTITION BY $2 ORDER BY $4
NULLS FIRST RANGE 1000:INTERVAL SECOND PRECEDING)])
+LogicalProject(c=[$2], EXPR$1=[COUNT($0) OVER (ORDER BY $4 NULLS FIRST ROWS 5
PRECEDING)])
+- LogicalTableScan(table=[[default_catalog, default_database, MyTable]])
]]>
</Resource>
<Resource name="optimized exec plan">
<![CDATA[
Calc(select=[c, w0$o0 AS $1])
-+- OverAggregate(partitionBy=[c], orderBy=[rowtime ASC], window=[ RANG BETWEEN
1000 PRECEDING AND CURRENT ROW], select=[a, c, rowtime, COUNT(a) AS w0$o0])
- +- Exchange(distribution=[hash[c]])
++- OverAggregate(orderBy=[rowtime ASC], window=[ ROWS BETWEEN 5 PRECEDING AND
CURRENT ROW], select=[a, c, rowtime, COUNT(a) AS w0$o0])
+ +- Exchange(distribution=[single])
+- Calc(select=[a, c, rowtime])
+- DataStreamScan(table=[[default_catalog, default_database,
MyTable]], fields=[a, b, c, proctime, rowtime])
]]>
</Resource>
</TestCase>
- <TestCase name="testRowTimeBoundedNonPartitionedRowsOver">
+ <TestCase name="testRowTimeBoundedPartitionedRangeOver">
<Resource name="sql">
<![CDATA[
SELECT c,
- COUNT(a) OVER (ORDER BY rowtime ROWS BETWEEN 5 PRECEDING AND CURRENT ROW)
-FROM MyTable
+ COUNT(a) OVER (PARTITION BY c ORDER BY rowtime
+ RANGE BETWEEN INTERVAL '1' SECOND PRECEDING AND CURRENT ROW)
+ FROM MyTable
]]>
</Resource>
<Resource name="ast">
<![CDATA[
-LogicalProject(c=[$2], EXPR$1=[COUNT($0) OVER (ORDER BY $4 NULLS FIRST ROWS 5
PRECEDING)])
+LogicalProject(c=[$2], EXPR$1=[COUNT($0) OVER (PARTITION BY $2 ORDER BY $4
NULLS FIRST RANGE 1000:INTERVAL SECOND PRECEDING)])
+- LogicalTableScan(table=[[default_catalog, default_database, MyTable]])
]]>
</Resource>
<Resource name="optimized exec plan">
<![CDATA[
Calc(select=[c, w0$o0 AS $1])
-+- OverAggregate(orderBy=[rowtime ASC], window=[ ROWS BETWEEN 5 PRECEDING AND
CURRENT ROW], select=[a, c, rowtime, COUNT(a) AS w0$o0])
- +- Exchange(distribution=[single])
++- OverAggregate(partitionBy=[c], orderBy=[rowtime ASC], window=[ RANG BETWEEN
1000 PRECEDING AND CURRENT ROW], select=[a, c, rowtime, COUNT(a) AS w0$o0])
Review Comment:
it behaves like that more than 5 years
https://github.com/apache/flink/blame/91aecc62f3c83956af4b3554f31ee2bbce057be7/flink-table/flink-table-planner/src/main/scala/org/apache/flink/table/planner/plan/utils/RelExplainUtil.scala#L608
should we change it? (not sure if it can lead to regression then...)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscr...@flink.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
