bvarghese1 commented on PR #29285:
URL: https://github.com/apache/flink/pull/29285#issuecomment-6022678898
> looks like I found regression (this passes in master and fails in PR's
branch)
>
> ```java
> class MapJoinProbeTest extends TableTestBase {
>
> private static final TypeInformation<?> INT_INFO =
BasicTypeInfo.INT_TYPE_INFO;
>
> @Test
> void probe() {
> BatchTableTestUtil util = batchTestUtil(TableConfig.getDefault());
> util.tableEnv()
> .getConfig()
> .set(
>
ExecutionConfigOptions.TABLE_EXEC_DISABLED_OPERATORS,
> "ShuffleHashJoin, BroadcastHashJoin");
> util.addTableSource(
> "Tiny",
> new TypeInformation<?>[] {INT_INFO, new
MapTypeInfo<>(INT_INFO, INT_INFO)},
> new String[] {"a", "m"},
> FlinkStatistic.builder().tableStats(new
TableStats(2L)).build());
> util.addTableSource(
> "Huge",
> new TypeInformation<?>[] {INT_INFO, new
MapTypeInfo<>(INT_INFO, INT_INFO)},
> new String[] {"d", "n"},
> FlinkStatistic.builder().tableStats(new
TableStats(1_000_000L)).build());
> util.tableEnv()
> .explainSql("SELECT a, d FROM Tiny, Huge
WHERE m = n");
> }
> }
> ```
Thanks @snuyanzin. I have added some type checking and for types that are
not sortable, NLJ still remains a valid fallback.
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
