[GitHub] [spark] kevincmchen commented on a change in pull request #33825: [SPARK-36568][SQL] Better FileScan statistics estimation

[GitBox <gi...@apache.org>]

kevincmchen commented on a change in pull request #33825:
URL: https://github.com/apache/spark/pull/33825#discussion_r695439272



##########
File path: sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/v2/FileScan.scala
##########
@@ -187,7 +189,10 @@ trait FileScan extends Scan
     new Statistics {
       override def sizeInBytes(): OptionalLong = {
         val compressionFactor = sparkSession.sessionState.conf.fileCompressionFactor
-        val size = (compressionFactor * fileIndex.sizeInBytes).toLong
+        val size = (compressionFactor * fileIndex.sizeInBytes /
+          (dataSchema.defaultSize + fileIndex.partitionSchema.defaultSize) *
+          (readDataSchema.defaultSize + readPartitionSchema.defaultSize)).toLong

Review comment:
       for parquet, which  is a columnar storage format, the  estimated size is much more correct; but for text, csv file,  the size which is computed by the old way, is more correct.




-- 
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: reviews-unsubscribe@spark.apache.org

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



---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org