[wp-trac] Re: [WordPress Trac] #9167: query_posts('meta_key=foo') returns duplicate posts

[WordPress Trac]

#9167: query_posts('meta_key=foo') returns duplicate posts
--------------------------+-------------------------------------------------
 Reporter:  scribu        |       Owner:  anonymous
     Type:  defect (bug)  |      Status:  new      
 Priority:  normal        |   Milestone:  2.7.2    
Component:  General       |     Version:           
 Severity:  normal        |    Keywords:           
--------------------------+-------------------------------------------------

Comment(by scribu):

 Replying to [comment:5 filosofo]:
 > Replying to [comment:2 scribu]:
 > > Can't this be solved by simply using "SELECT DISTINCT(wp_posts.ID)..."
 on all queries?
 >
 > No, that won't help for queries like this, because each ''row'' is
 already distinct.  Besides, DISTINCT clauses usually have a performance
 penalty, and this is the core query of WP.

 If you write ''SELECT DISTINCT(*)'', no, it won't; but if you write
 ''SELECT DISTINCT(wp_posts.ID), wp_posts.post_author, etc.'' it will work.
 The only problem is that you couldn't ''SELECT wp_posts.*'' anymore.

 I think a subquery also has a performance penalty. Nothing you can do
 about that.

 By the way, I see that subqueries are already used in WP 2.7.1 (in
 query.php):

 {{{
 $whichcat .= " AND $wpdb->posts.ID NOT IN ( SELECT tr.object_id FROM
 $wpdb->term_relationships AS tr INNER JOIN $wpdb->term_taxonomy AS tt ON
 tr.term_taxonomy_id = tt.term_taxonomy_id WHERE tt.taxonomy = 'post_tag'
 AND tt.term_id IN ($tag_string) )";
 }}}

-- 
Ticket URL: <http://core.trac.wordpress.org/ticket/9167#comment:6>
WordPress Trac <http://trac.wordpress.org/>
WordPress blogging software