Skip to content

[SPARK-17182][SQL] Mark Collect as non-deterministic #14749

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
liancheng wants to merge 1 commit into from
Closed

[SPARK-17182][SQL] Mark Collect as non-deterministic #14749

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions ...catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/aggregate/collect.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -54,6 +54,10 @@ abstract class Collect extends ImperativeAggregate {

override def inputAggBufferAttributes: Seq[AttributeReference] = Nil

// Both `CollectList` and `CollectSet` are non-deterministic since their results depend on the
// actual order of input rows.
override def deterministic: Boolean = false
Copy link
Contributor

@hvanhovell hvanhovell Aug 22, 2016

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is collect_set non deterministic? It is backed by a HashSet, and the way elements are iterated over does not rely on the input order.

Copy link
Contributor

@hvanhovell hvanhovell Aug 22, 2016

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

NVM. Actually iteration over a set is non-deterministic for values in the same hash bucket.

Copy link
Contributor Author

@liancheng liancheng Aug 23, 2016

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yea. Also, there's no set type in Spark SQL, thus the result set is actually just a list without duplicated elements.


protected[this] val buffer: Growable[Any] with Iterable[Any]

override def initialize(b: MutableRow): Unit = {
Expand Down