Github user dongjoon-hyun commented on a diff in the pull request:
https://github.com/apache/spark/pull/14132#discussion_r71463733
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/catalyst/SQLBuilder.scala ---
@@ -356,8 +372,14 @@ class SQLBuilder(logicalPlan: LogicalPlan) extends
Logging {
}
}
+ val broadcastHint = project match {
+ case p @ Project(projectList, Hint("BROADCAST", tables, child)) =>
+ if (tables.nonEmpty) s"/*+ MAPJOIN(${tables.mkString(", ")}) */"
else ""
+ case _ => ""
+ }
build(
"SELECT",
+ broadcastHint,
--- End diff --
It's the result of Window test query. For the Windows query, there were
nested Projects.
```
test("broadcast hint with window") {
checkSQL(
"""
|SELECT /*+ MAPJOIN(parquet_t1) */
| x.key, MAX(y.key) OVER (PARTITION BY x.key % 5 ORDER BY
x.key)
|FROM parquet_t1 x JOIN parquet_t1 y ON x.key = y.key
""".stripMargin,
"broadcast_hint_window")
}
```
I had the same feeling why some "SELECT" doesn't happen.
After https://issues.apache.org/jira/browse/SPARK-16576 , I think this kind
of weirdness will be reduced.
---
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.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
