srowen commented on a change in pull request #26415: [SPARK-18409][ML] LSH
approxNearestNeighbors should use approxQuantile instead of sort
URL: https://github.com/apache/spark/pull/26415#discussion_r348906242
##########
File path: mllib/src/main/scala/org/apache/spark/ml/feature/LSH.scala
##########
@@ -112,7 +112,9 @@ private[ml] abstract class LSHModel[T <: LSHModel[T]]
numNearestNeighbors: Int,
singleProbe: Boolean,
distCol: String): Dataset[_] = {
- require(numNearestNeighbors > 0, "The number of nearest neighbors cannot
be less than 1")
+ val count = dataset.count()
Review comment:
It implies it _could_ have fewer, yeah. This would be the only case to
return fewer, yes, and you'd just return the input. I don't think singleProbe
is used (and could be removed). But the count() is still needed below in any
event.
I mean, if the caller cares, they can check count() vs number of neighbors
before calling it a bunch of times anyway.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
