[
https://issues.apache.org/jira/browse/HIVE-15537?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15797340#comment-15797340
]
Ferdinand Xu commented on HIVE-15537:
-------------------------------------
Do we need to keep the original unit test over the following fields to get rid
of regression?
{noformat}
s3 struct<f12: array<struct<f13:string, f14:int>>>,
s4 map<string, struct<f15:int>>
{noformat}
> Nested column pruning: fix issue when selecting struct field from array/map
> element (part 2)
> --------------------------------------------------------------------------------------------
>
> Key: HIVE-15537
> URL: https://issues.apache.org/jira/browse/HIVE-15537
> Project: Hive
> Issue Type: Sub-task
> Components: Logical Optimizer
> Affects Versions: 2.2.0
> Reporter: Chao Sun
> Assignee: Chao Sun
> Attachments: HIVE-15537.1.patch
>
>
> HIVE-15507 only addresses the issue of
> {code}
> SELECT arr[0].f FROM tbl
> {code}
> However, it didn't handle:
> {code}
> SELECT arr[0].f.g FROM tbl
> {code}
> In this case the current code will generate a path {{arr.g}}, which is wrong.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
