[
https://issues.apache.org/jira/browse/HIVE-7221?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14029401#comment-14029401
]
Vaibhav Gumashta commented on HIVE-7221:
----------------------------------------
Actually there is an option (which wasn't documented earlier) which lets you
print output incrementally: beeline --incremental=true. I think we should have
that true by default. Will create another jira for that.
> Beeline buffers the entire output file in memory before writing to stdout
> -------------------------------------------------------------------------
>
> Key: HIVE-7221
> URL: https://issues.apache.org/jira/browse/HIVE-7221
> Project: Hive
> Issue Type: Bug
> Components: Clients, JDBC
> Reporter: Vaibhav Gumashta
> Fix For: 0.13.0
>
>
> It seems beeline does not write to stdout till it reads the entire output
> relation. This can cause OOM and should be fixed. Beeline should only buffer
> a small number of row batches.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
