Github user kiszk commented on a diff in the pull request:
https://github.com/apache/spark/pull/17122#discussion_r103923150
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/WholeStageCodegenExec.scala
---
@@ -206,6 +206,18 @@ trait CodegenSupport extends SparkPlan {
def doConsume(ctx: CodegenContext, input: Seq[ExprCode], row: ExprCode):
String = {
throw new UnsupportedOperationException
}
+
+ /**
+ * for optimization to suppress shouldStop() in a loop of
WholeStageCodegen
+ *
+ * isShouldStopRequired: require to insert shouldStop() into the loop if
true
+ */
+ def isShouldStopRequired: Boolean = {
+ shouldStopRequired || (this.parent != null &&
this.parent.isShouldStopRequired)
+ }
+
+ // set true if doConsume() inserts append() method that requires
shouldStop() in the loop
+ protected var shouldStopRequired: Boolean = false
--- End diff --
@hvanhovell We can do it technically. It looks simple. On the other hand,
we have to maintain an immutable var carefully by investigating whether
`append()` is required. In particular, we would add a new
CodegenSupport-related class.
In contrast, the current approach is easy to maintain the var. When we
would add a new CodegenSupport-related class, it is unnecessary to carefully
investigate it.
This is a trade-off between simplicity and maintainability. What do you
think?
---
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]
