Herman van Hovell created SPARK-14950:
-----------------------------------------
Summary: Anti Join is broken when using BroadcastHashJoin's unique
key path
Key: SPARK-14950
URL: https://issues.apache.org/jira/browse/SPARK-14950
Project: Spark
Issue Type: Bug
Components: SQL
Reporter: Herman van Hovell
Assignee: Herman van Hovell
Anti Join is broken when using BroadcastHashJoin's unique key path. For example:
{noformat}
test("left anti join") {
val df = Seq(1, 3).toDF("int")
val df2 = Seq(1, 5).toDF("int2")
checkAnswer(
df.as("a").join(df2.as("b"), $"a.int" === $"b.int2", "left_anti"),
Row(3) :: Nil)
}
{noformat}
Returns {{df}} instead of an empty dataframe. Anti Join now behaves like a Semi
Join.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
