[ 
https://issues.apache.org/jira/browse/HIVE-15743?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15840955#comment-15840955
 ] 

Matt McCline commented on HIVE-15743:
-------------------------------------

Wow that is a high percentage!

Ya, I looked at some of those Java runtime classes recently (when I was doing 
FastDecimal) and it is doable.  All numeric related characters are UTF-8.   
There is considerable magic code around exponents, etc in them though but 
because of that I doubt few people are brave enough to change it.  So, our 
transformed version would be "relatively safe".

> vectorized text parsing: speed up double parse
> ----------------------------------------------
>
>                 Key: HIVE-15743
>                 URL: https://issues.apache.org/jira/browse/HIVE-15743
>             Project: Hive
>          Issue Type: Bug
>            Reporter: Sergey Shelukhin
>         Attachments: tpch-without.png
>
>
> {noformat}
> Double.parseDouble(
>                 new String(bytes, fieldStart, fieldLength, 
> StandardCharsets.UTF_8));{noformat}
> This takes ~25% of the query time in some cases.



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

Reply via email to