[jira] [Comment Edited] (HIVE-13098) Add a strict check for when the decimal gets converted to null due to insufficient width

["Sergey Shelukhin (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HIVE-13098?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15534285#comment-15534285 ] 

Sergey Shelukhin edited comment on HIVE-13098 at 9/29/16 10:47 PM:
-------------------------------------------------------------------

Upon looking further on OI path I don't think it's possible to propagate it there without major changes, in fact OI-related parts of this patch are not valid, since OIs are assumed to be stateless and are cached process-wide, ditto for TypeInfo-s. There are lots of static method paths accessing those...
I think I might scrape a lot of the patch and add a globally accessible static that would have to be initialize on CLI/HS2/task startup.. The only exception would be write path that happens outside of Hive services... 

This will reduce size of the patch a lot (but also make it a global setting not modifiable per query...)

Update: another alternative would be a (TADA!) threadlocal.
We could set it at compile time and change the patch to have only compile paths use it, whereas runtime paths would use the fields in OIs and fns that compile populates.

[~ashutoshc] [~hagleitn] [~jdere] opinions?


was (Author: sershe):
Upon looking further on OI path I don't think it's possible to propagate it there without major changes, in fact OI-related parts of this patch are not valid, since OIs are assumed to be stateless and are cached process-wide, ditto for TypeInfo-s. There are lots of static method paths accessing those...
I think I might scrape a lot of the patch and add a globally accessible static that would have to be initialize on CLI/HS2/task startup.. The only exception would be write path that happens outside of Hive services... 

This will reduce size of the patch a lot (but also make it a global setting not modifiable per query...)

[~ashutoshc] [~hagleitn] [~jdere] opinions?

> Add a strict check for when the decimal gets converted to null due to insufficient width
> ----------------------------------------------------------------------------------------
>
>                 Key: HIVE-13098
>                 URL: https://issues.apache.org/jira/browse/HIVE-13098
>             Project: Hive
>          Issue Type: Bug
>            Reporter: Sergey Shelukhin
>            Assignee: Sergey Shelukhin
>         Attachments: HIVE-13098.WIP.patch, HIVE-13098.WIP2.patch
>
>
> When e.g. 999999 is selected as decimal(5,0), the result is null. This can be problematic, esp. if the data is written to a table and lost without the user realizing it. There should be an option to error out in such cases instead; it should probably be on by default and the error message should instruct the user on how to disable it.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)