[jira] [Commented] (DRILL-7771) Order by and limit transactions in RDBMS cannot be pushdown

["ASF GitHub Bot (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/DRILL-7771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17241845#comment-17241845 ] 

ASF GitHub Bot commented on DRILL-7771:
---------------------------------------

vvysotskyi opened a new pull request #2119:
URL: https://github.com/apache/drill/pull/2119


   # [DRILL-7771](https://issues.apache.org/jira/browse/DRILL-7771): Order by and limit transactions in RDBMS cannot be pushdown
   
   ## Description
   - Introduced `DrillJdbcSort` to have the cost for limit queries that correspond to Drill's limit cost.
   - Provided the ability to apply some JDBC rules after the logical planning phase. All other JDBC rules will be addressed in https://issues.apache.org/jira/browse/CALCITE-4423.
   - Added unit tests
   
   ## Documentation
   NA
   
   ## Testing
   Added unit tests for the cases that were broken earlier.
   


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


> Order by and limit transactions in RDBMS cannot be pushdown
> -----------------------------------------------------------
>
>                 Key: DRILL-7771
>                 URL: https://issues.apache.org/jira/browse/DRILL-7771
>             Project: Apache Drill
>          Issue Type: Bug
>          Components: Query Planning & Optimization
>    Affects Versions: 1.17.0
>            Reporter: Yusuf
>            Assignee: Vova Vysotskyi
>            Priority: Critical
>         Attachments: Physical Plan.PNG, query.PNG
>
>
> hi, 
> When using ORDER BY, LIMIT and OFFSET in large tables in RDBMS, it draws the entire table on DRILL and it occurs very slow or heap memory. When we look at the query plan, instead of having ORDER BY, LIMIT and OFFSET features done by RDBMS, DRILL does it by pulling all the data on it. It is very critical to transmit these transactions to the database.
> ISSUE # 7490
> https://issues.apache.org/jira/browse/drıll-7490
> The LIMIT pushdown problem was communicated as resolved but there was a missing solution. I wrote the necessary comment in the comment section.
> Regardless of the number of columns selected, regardless of the number of joins and tables, pushdown should be done in RDBMS



--
This message was sent by Atlassian Jira
(v8.3.4#803005)