[jira] [Updated] (ARROW-14389) [C++][Gandiva] Fix performance bug with LIKE expressions with reserved characters

["Krisztian Szucs (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/ARROW-14389?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Krisztian Szucs updated ARROW-14389:
------------------------------------
    Fix Version/s: 8.0.0
                       (was: 7.0.0)

> [C++][Gandiva] Fix performance bug with LIKE expressions with reserved characters
> ---------------------------------------------------------------------------------
>
>                 Key: ARROW-14389
>                 URL: https://issues.apache.org/jira/browse/ARROW-14389
>             Project: Apache Arrow
>          Issue Type: Bug
>          Components: C++ - Gandiva
>            Reporter: João Victor Huguenin
>            Assignee: João Victor Huguenin
>            Priority: Major
>              Labels: pull-request-available
>             Fix For: 8.0.0
>
>          Time Spent: 2h 50m
>  Remaining Estimate: 0h
>
> Seeing dramatic differences in Gandiva processing time during filter operations for queries with a LIKE filter on a character field with a reserved character (-,/& etc). The version with a reserved character in the string is considerably slower than a similar filter without a reserved character. For patterns like %abc% and %ab-c%, the latter wasn't being optimized to become an is_substr expression because of the regex used to identify those cases.



--
This message was sent by Atlassian Jira
(v8.20.7#820007)