[
https://issues.apache.org/jira/browse/HIVE-6632?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13931978#comment-13931978
]
Yin Huai commented on HIVE-6632:
--------------------------------
Does Parquet have the same issue?
> ORC should be able to only read needed fields in a complex column
> -----------------------------------------------------------------
>
> Key: HIVE-6632
> URL: https://issues.apache.org/jira/browse/HIVE-6632
> Project: Hive
> Issue Type: Improvement
> Reporter: Yin Huai
>
> Currently, we use a string of ids to record needed columns. However, this
> string cannot record needed fields of a complex column. Although ORC
> decomposes a complex column to multiple sub-columns, it has to load the
> entire complex column if only a single field of this complex column is needed.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
