mridulm commented on code in PR #38428:
URL: https://github.com/apache/spark/pull/38428#discussion_r1090263540
##########
core/src/main/scala/org/apache/spark/util/collection/ExternalAppendOnlyMap.scala:
##########
@@ -504,44 +505,31 @@ class ExternalAppendOnlyMap[K, V, C](
* If no more pairs are left, return null.
*/
private def readNextItem(): (K, C) = {
- try {
- val k = deserializeStream.readKey().asInstanceOf[K]
- val c = deserializeStream.readValue().asInstanceOf[C]
- val item = (k, c)
- objectsRead += 1
- if (objectsRead == serializerBatchSize) {
- objectsRead = 0
- deserializeStream = nextBatchStream()
- }
- item
- } catch {
- case e: EOFException =>
- cleanup()
- null
+ val next = batchIterator.next()
+ objectsRead += 1
+ if (objectsRead == serializerBatchSize) {
+ objectsRead = 0
+ batchIterator = nextBatchIterator()
}
+ next
}
override def hasNext: Boolean = {
- if (nextItem == null) {
- if (deserializeStream == null) {
- // In case of deserializeStream has not been initialized
- deserializeStream = nextBatchStream()
- if (deserializeStream == null) {
- return false
- }
+ if (batchIterator == null) {
+ // In case of batchIterator has not been initialized
+ batchIterator = nextBatchIterator()
+ if (batchIterator == null) {
+ return false
}
- nextItem = readNextItem()
}
- nextItem != null
+ batchIterator.hasNext
Review Comment:
Review Note: this change is preserving the behavior which exists - but as a
general construct, if the returned `batchIterator` ends up being empty - we
wont go to the next batch iterator.
This is not expected to be an issue though.
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
