HyukjinKwon commented on a change in pull request #23830:
[SPARK-26935][SQL]Skip DataFrameReader's CSV first line scan when not used
URL: https://github.com/apache/spark/pull/23830#discussion_r258362483
##########
File path: sql/core/src/main/scala/org/apache/spark/sql/DataFrameReader.scala
##########
@@ -508,7 +508,12 @@ class DataFrameReader private[sql](sparkSession:
SparkSession) extends Logging {
sparkSession.sessionState.conf.sessionLocalTimeZone)
val filteredLines: Dataset[String] =
CSVUtils.filterCommentAndEmpty(csvDataset, parsedOptions)
- val maybeFirstLine: Option[String] = filteredLines.take(1).headOption
+ val maybeFirstLine: Option[String] =
+ if (userSpecifiedSchema.isEmpty || parsedOptions.headerFlag) {
+ filteredLines.take(1).headOption
+ } else {
+ None
Review comment:
Two things to note ..
1. Can we somehow put this logic into `CSVHeaderChecker` and
`TextInputCSVDataSource`? The problem is, first line is used in both places but
it should be collected only one time.
2. If that's tricky, let's add some comments here. CSV code is quite
complicated and I would like to avoid to add such fixes as possible as I can.
If this fix is inevitable, let's at least add some comments.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on 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]
