[
https://issues.apache.org/jira/browse/HIVE-6714?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13942390#comment-13942390
]
Gunther Hagleitner commented on HIVE-6714:
------------------------------------------
I think a cleaner way would be to eliminate duplicate keys during parse stage.
The deserialized structs would only have each key once. That way you don't have
to carry both actual (deduped) and original size. Is that possible?
> Fix getMapSize() of LazyMap
> ---------------------------
>
> Key: HIVE-6714
> URL: https://issues.apache.org/jira/browse/HIVE-6714
> Project: Hive
> Issue Type: Sub-task
> Components: Serializers/Deserializers
> Affects Versions: 0.13.0, 0.14.0
> Reporter: Prasanth J
> Assignee: Prasanth J
> Fix For: 0.13.0, 0.14.0
>
> Attachments: HIVE-6714.1.patch
>
>
> HIVE-6707 fixed an issue with lazy maps. Lazy maps scans for keys using key
> separators and maintains a counter for the number of keys i.e; map size.
> However keys may include duplicates in which returning the number of keys for
> map size will be wrong as after deserialization the actual map will not have
> duplicate keys.
> This is a subtask of HIVE-6707 to fix the getMapSize() api.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
