You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@celeborn.apache.org by "zhouyifan279 (via GitHub)" <gi...@apache.org> on 2023/08/25 11:30:04 UTC

[GitHub] [incubator-celeborn] zhouyifan279 commented on a diff in pull request #1843: [CELEBORN-919] Move Columnar Shuffle code into an individual module

zhouyifan279 commented on code in PR #1843:
URL: https://github.com/apache/incubator-celeborn/pull/1843#discussion_r1305540695


##########
client-spark/spark-3/src/main/scala/org/apache/spark/shuffle/celeborn/CelebornShuffleReader.scala:
##########
@@ -50,21 +49,7 @@ class CelebornShuffleReader[K, C](
 
   override def read(): Iterator[Product2[K, C]] = {
 
-    var serializerInstance = dep.serializer.newInstance()
-    if (conf.columnarShuffleEnabled) {
-      val schema = SparkUtils.getSchema(dep)
-      if (CelebornBatchBuilder.supportsColumnarType(
-          schema)) {
-        val dataSize = SparkUtils.getDataSize(
-          dep.serializer.asInstanceOf[UnsafeRowSerializer])
-        serializerInstance = new CelebornColumnarBatchSerializer(
-          schema,
-          conf.columnarShuffleBatchSize,
-          conf.columnarShuffleDictionaryEnabled,
-          conf.columnarShuffleOffHeapEnabled,
-          dataSize).newInstance()
-      }
-    }
+    val serializerInstance = newSerializerInstance(dep)

Review Comment:
   Gluten Columnar Shuffle is different from Celeborn Columnar Shuffle.
   It is clean and won't stop us to upgrade spark dependency version.
   Let's leave it here for now.



-- 
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: issues-unsubscribe@celeborn.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org