Github user fhueske commented on a diff in the pull request:
https://github.com/apache/flink/pull/3673#discussion_r115785828
--- Diff:
flink-libraries/flink-table/src/test/scala/org/apache/flink/table/api/scala/batch/sql/DataSetSingleRowJoinTest.scala
---
@@ -187,9 +187,106 @@ class SingleRowJoinTest extends TableTestBase {
),
term("where", "AND(<(a1, b1)", "=(a2, b2))"),
term("join", "a1", "a2", "b1", "b2"),
- term("joinType", "NestedLoopJoin")
+ term("joinType", "NestedLoopInnerJoin")
)
util.verifySql(query, expected)
}
+
+ @Test
+ def testSingleRowJoinLeftOuterJoin(): Unit = {
+ val util = batchTestUtil()
+ util.addTable[(Long, Int)]("A", 'a1, 'a2)
+ util.addTable[(Int, Int)]("B", 'b1, 'b2)
+
+ val queryLeftJoin =
+ "SELECT a2 FROM A " +
+ "LEFT JOIN " +
+ "(SELECT COUNT(*) AS cnt FROM B) " +
+ "AS x " +
+ "ON a1 = cnt"
+
+ val expected =
+ unaryNode(
+ "DataSetCalc",
+ unaryNode(
+ "DataSetSingleRowJoin",
+ batchTableNode(0),
+ term("where", "=(a1, cnt)"),
+ term("join", "a1", "a2", "cnt"),
+ term("joinType", "NestedLoopLeftJoin")
+ ),
+ term("select", "a2")
+ ) + "\n" +
+ unaryNode(
+ "DataSetAggregate",
+ unaryNode(
+ "DataSetUnion",
+ unaryNode(
+ "DataSetValues",
+ unaryNode(
+ "DataSetCalc",
+ batchTableNode(1),
+ term("select", "0 AS $f0")),
+ tuples(List(null)), term("values", "$f0")
+ ),
+ term("union", "$f0")
+ ),
+ term("select", "COUNT(*) AS cnt")
+ )
+
+ util.verifySql(queryLeftJoin, expected)
+ }
+
+ @Test
+ def testSingleRowJoinInnerJoin(): Unit = {
+ val util = batchTestUtil()
+ util.addTable[(Int, Int)]("A", 'a1, 'a2)
+ val query =
+ "SELECT a2, sum(a1) " +
+ "FROM A " +
+ "GROUP BY a2 " +
+ "HAVING sum(a1) > (SELECT sum(a1) * 0.1 FROM A)"
+
+ val expected =
+ unaryNode(
+ "DataSetCalc",
+ unaryNode(
+ "DataSetSingleRowJoin",
+ unaryNode(
+ "DataSetAggregate",
+ batchTableNode(0),
+ term("groupBy", "a2"),
+ term("select", "a2", "SUM(a1) AS EXPR$1")
+ ),
+ term("where", ">(EXPR$1, EXPR$0)"),
+ term("join", "a2", "EXPR$1", "EXPR$0"),
+ term("joinType", "NestedLoopInnerJoin")
+ ),
+ term("select", "a2", "EXPR$1")
+ ) + "\n" +
+ unaryNode(
+ "DataSetCalc",
+ unaryNode(
+ "DataSetAggregate",
+ unaryNode(
+ "DataSetUnion",
+ unaryNode(
+ "DataSetValues",
+ unaryNode(
+ "DataSetCalc",
+ batchTableNode(0),
+ term("select", "a1")
+ ),
+ tuples(List(null)), term("values", "a1")
+ ),
+ term("union", "a1")
+ ),
+ term("select", "SUM(a1) AS $f0")
+ ),
+ term("select", "*($f0, 0.1) AS EXPR$0")
+ )
+
+ util.verifySql(query, expected)
+ }
--- End diff --
I add the following additional unit tests here:
- left join with inequality join predicate
- right join with equality join predicate
- right join with inequality join predicate
You will find that the left/right joins with inequality predicates cannot
be translated because of the too strict condition in
`FlinkLogicalJoinConverter`. This condition needs to be adapted to accept
single row outer joins with inequality predicates.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
