Github user HeartSaVioR commented on a diff in the pull request:
https://github.com/apache/spark/pull/22952#discussion_r231695749
--- Diff: docs/structured-streaming-programming-guide.md ---
@@ -530,6 +530,8 @@ Here are the details of all the sources in Spark.
"s3://a/dataset.txt"<br/>
"s3n://a/b/dataset.txt"<br/>
"s3a://a/b/c/dataset.txt"<br/>
+ <br/>
+ <code>renameCompletedFiles</code>: whether to rename completed
files in previous batch (default: false). If the option is enabled, input file
will be renamed with additional postfix "_COMPLETED_". This is useful to clean
up old input files to save space in storage.
--- End diff --
@dongjoon-hyun
For Storm, it renames input file twice, 1. in process 2. completed
(actually it is not a rename, but move to archive directory). HDFS spout is
created at 2015 which I don't expect there's deep consideration on cloud
storage.
For Flink I have no idea, I'll explore how they handle it.
I think the feature is just an essential thing in ETL situation: a comment
in JIRA clearly shows why the feature is needed.
https://issues.apache.org/jira/browse/SPARK-20568?focusedCommentId=16356929&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-16356929
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
