[
https://issues.apache.org/jira/browse/HIVE-13178?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15196415#comment-15196415
]
Prasanth Jayachandran commented on HIVE-13178:
----------------------------------------------
I haven't gone through core changes. Left some initial comments. Main concerns
is we are bringing in ObjectInspector back into tree readers which will make it
difficult to separate ORC out of hive. If this feature is targeted to be
supported inside of orc then these object inspectors should be replaced by
TypeDescriptors. Also it will good to subclass TreeReaderFactory to handle type
conversions.
> Enhance ORC Schema Evolution to handle more standard data type conversions
> --------------------------------------------------------------------------
>
> Key: HIVE-13178
> URL: https://issues.apache.org/jira/browse/HIVE-13178
> Project: Hive
> Issue Type: Bug
> Components: Hive, ORC
> Reporter: Matt McCline
> Assignee: Matt McCline
> Priority: Critical
> Attachments: HIVE-13178.01.patch, HIVE-13178.02.patch,
> HIVE-13178.03.patch
>
>
> Currently, SHORT -> INT -> BIGINT is supported.
> Handle ORC data type conversions permitted by Implicit conversion allowed by
> TypeIntoUtils.implicitConvertible method.
> * STRING_GROUP -> DOUBLE
> * STRING_GROUP -> DECIMAL
> * DATE_GROUP -> STRING
> * NUMERIC_GROUP -> STRING
> * STRING_GROUP -> STRING_GROUP
> *
> * // Upward from "lower" type to "higher" numeric type:
> * BYTE -> SHORT -> INT -> BIGINT -> FLOAT -> DOUBLE -> DECIMAL
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
