turboFei commented on a change in pull request #26339:
[SPARK-27194][SPARK-29302][SQL] Fix the issue that for dynamic partition
overwrite a task would conflict with its speculative task
URL: https://github.com/apache/spark/pull/26339#discussion_r355823338
##########
File path:
core/src/main/scala/org/apache/spark/internal/io/HadoopMapReduceCommitProtocol.scala
##########
@@ -236,13 +271,23 @@ class HadoopMapReduceCommitProtocol(
committer.setupTask(taskContext)
addedAbsPathFiles = mutable.Map[String, String]()
partitionPaths = mutable.Set[String]()
+ stagingTaskFiles = mutable.Set[Path]()
}
override def commitTask(taskContext: TaskAttemptContext): TaskCommitMessage
= {
val attemptId = taskContext.getTaskAttemptID
logTrace(s"Commit task ${attemptId}")
SparkHadoopMapRedUtil.commitTask(
committer, taskContext, attemptId.getJobID.getId,
attemptId.getTaskID.getId)
+ if (dynamicPartitionOverwrite) {
+ val fs = stagingDir.getFileSystem(taskContext.getConfiguration)
+ stagingTaskFiles.foreach { stagingTaskFile =>
+ val fileName = stagingTaskFile.getName
+ val taskPartitionPath = getPartitionPath(stagingTaskFile)
+ val destFile = new Path(new Path(stagingDir, taskPartitionPath),
fileName)
+ fs.rename(stagingTaskFile, destFile)
Review comment:
Thanks for your suggestion, but I think we can not do this.
For dynamic partition overwrite, each task may write data to many partition
paths.
If we just rename a folder, it would overwrite other task's output.
----------------------------------------------------------------
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]
