You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Vladimir Ozerov (JIRA)" <ji...@apache.org> on 2019/02/14 08:50:00 UTC

[jira] [Comment Edited] (IGNITE-6877) Improve behaviour for non-correlated subqueries

    [ https://issues.apache.org/jira/browse/IGNITE-6877?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16768006#comment-16768006 ] 

Vladimir Ozerov edited comment on IGNITE-6877 at 2/14/19 8:49 AM:
------------------------------------------------------------------

Closing due to long inactivity. Planner logic is to be redesigned soon, so tickets will be created as needed.


was (Author: vozerov):
Closing due to long inactivity. Planned logic is to be redesigned soon, so tickets will be created as needed.

> Improve behaviour for non-correlated subqueries
> -----------------------------------------------
>
>                 Key: IGNITE-6877
>                 URL: https://issues.apache.org/jira/browse/IGNITE-6877
>             Project: Ignite
>          Issue Type: Improvement
>          Components: sql
>    Affects Versions: 2.3
>            Reporter: Alin Andrei Corodescu
>            Priority: Minor
>
> Ignite behaves poorly in terms of performance when given queries which contain IN or = operators followed by a non-correlated subquery. My guess is that the query is actually run for each row of the table in order to test the condition. 
> A possible solution is to store the results of the subquery in a temporary table, and use it from there.
> A workaround at the moment is to use joins instead of IN or = operators, but this makes the query much more complicated.
> Example:
> SELECT name
> FROM Employees
> WHERE age IN (SELECT age FROM Customers)
> The performance of this query is very slow for small amounts of data (about 20 seconds for 4000 rows in each table last time I tried)



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)