[
https://issues.apache.org/jira/browse/FLINK-5452?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15820585#comment-15820585
]
ASF GitHub Bot commented on FLINK-5452:
---------------------------------------
Github user fhueske commented on a diff in the pull request:
https://github.com/apache/flink/pull/3095#discussion_r95753596
--- Diff:
flink-libraries/flink-table/src/test/scala/org/apache/flink/table/api/scala/batch/sql/SortITCase.scala
---
@@ -55,14 +55,8 @@ class SortITCase(
tEnv.registerDataSet("MyTable", ds)
val expected = sortExpectedly(tupleDataSetStrings)
- val results = tEnv.sql(sqlQuery).toDataSet[Row].mapPartition(rows =>
Seq(rows.toSeq)).collect()
-
- val result = results
- .filterNot(_.isEmpty)
- .sortBy(_.head)(Ordering.by(f=> f.toString))
--- End diff --
The problem here is the string conversion which results in a
lexicographical order ("21,....." sorts before "5,...").
We should change this to `.sortBy(_.head)(Ordering.by(r => (r.getField(0),
r.getField(1))))`
> Make table unit tests pass under cluster mode
> ---------------------------------------------
>
> Key: FLINK-5452
> URL: https://issues.apache.org/jira/browse/FLINK-5452
> Project: Flink
> Issue Type: Improvement
> Components: Table API & SQL
> Reporter: Kurt Young
> Assignee: Kurt Young
>
> Currently if we change the test execution mode to
> {{TestExecutionMode.CLUSTER}} in {{TableProgramsTestBase}}, some cases will
> fail. Need to figure out whether it's the case design problem or there are
> some bugs.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
