[
https://issues.apache.org/jira/browse/HIVE-11316?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14637541#comment-14637541
]
Eugene Koifman commented on HIVE-11316:
---------------------------------------
My concern was with the modification of the existing toStringTree() in a
dangerous way.
patch 3 keeps toStringTree() as is and adds a new "optimized" method.
[~hsubramaniyan] and I just discussed and there is a best of both worlds
approach.
ASTNode only has 5 or 6 (inherited) methods that allow tree modification. We
could overload each one to set a flag that says cached "toString" needs to be
recomputed. This way no one even has to know about caching.
[~jcamachorodriguez], does this seem reasonable?
> Use datastructure that doesnt duplicate any part of string for
> ASTNode::toStringTree()
> --------------------------------------------------------------------------------------
>
> Key: HIVE-11316
> URL: https://issues.apache.org/jira/browse/HIVE-11316
> Project: Hive
> Issue Type: Bug
> Reporter: Hari Sankar Sivarama Subramaniyan
> Assignee: Hari Sankar Sivarama Subramaniyan
> Attachments: HIVE-11316-branch-1.0.patch,
> HIVE-11316-branch-1.2.patch, HIVE-11316.1.patch, HIVE-11316.2.patch,
> HIVE-11316.3.patch
>
>
> HIVE-11281 uses an approach to memoize toStringTree() for ASTNode. This jira
> is suppose to alter the string memoization to use a different data structure
> that doesn't duplicate any part of the string so that we do not run into OOM.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
