[jira] [Commented] (SPARK-31588) merge small files may need more common setting

2020-05-12 Thread philipse (Jira)
[ https://issues.apache.org/jira/browse/SPARK-31588?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17105530#comment-17105530 ] philipse commented on SPARK-31588: -- Thanks Hyukjin for your advice , i will reconsider it.

[jira] [Commented] (SPARK-31588) merge small files may need more common setting

2020-05-09 Thread Hyukjin Kwon (Jira)
[ https://issues.apache.org/jira/browse/SPARK-31588?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17103658#comment-17103658 ] Hyukjin Kwon commented on SPARK-31588: -- I can't completely get the point of the physical size and

[jira] [Commented] (SPARK-31588) merge small files may need more common setting

2020-05-08 Thread philipse (Jira)
[ https://issues.apache.org/jira/browse/SPARK-31588?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17102414#comment-17102414 ] philipse commented on SPARK-31588: -- yes, the block size can be controlled in HDFS.i mean we just take

[jira] [Commented] (SPARK-31588) merge small files may need more common setting

2020-05-07 Thread Hyukjin Kwon (Jira)
[ https://issues.apache.org/jira/browse/SPARK-31588?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17102195#comment-17102195 ] Hyukjin Kwon commented on SPARK-31588: -- the repartition won't set the hard limit on the size. You

[jira] [Commented] (SPARK-31588) merge small files may need more common setting

2020-05-07 Thread philipse (Jira)
[ https://issues.apache.org/jira/browse/SPARK-31588?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17101775#comment-17101775 ] philipse commented on SPARK-31588: -- For example: if we have output 3 files,size as 10M,50M,200M,the

[jira] [Commented] (SPARK-31588) merge small files may need more common setting

2020-05-02 Thread Hyukjin Kwon (Jira)
[ https://issues.apache.org/jira/browse/SPARK-31588?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17098210#comment-17098210 ] Hyukjin Kwon commented on SPARK-31588: -- There are many other workarounds already. Can you show a