besoin d'aide pour optimiser wordpress meta

Je suis tombé sur ce problème et j'ai réalisé que le problème était avec tous les INNER JOINS générés par WordPress. J'ai reçu la requête brute de WordPress :

SELECT   wp_posts.* FROM wp_posts  
INNER JOIN wp_term_relationships ON (wp_posts.ID = wp_term_relationships.object_id) 
INNER JOIN wp_postmeta ON (wp_posts.ID = wp_postmeta.post_id)
INNER JOIN wp_postmeta AS mt1 ON (wp_posts.ID = mt1.post_id)
INNER JOIN wp_postmeta AS mt2 ON (wp_posts.ID = mt2.post_id)
INNER JOIN wp_postmeta AS mt3 ON (wp_posts.ID = mt3.post_id)
INNER JOIN wp_postmeta AS mt4 ON (wp_posts.ID = mt4.post_id)
INNER JOIN wp_postmeta AS mt5 ON (wp_posts.ID = mt5.post_id) WHERE 1=1  AND ( wp_term_relationships.term_taxonomy_id IN (7,11,12,13,14,15) AND wp_posts.ID NOT IN (
                    SELECT object_id
                    FROM wp_term_relationships
                    WHERE term_taxonomy_id IN (10)
                ) ) AND wp_posts.post_type IN ('post', 'page', 'attachment', 'product', 'store_page') AND (wp_posts.post_status = 'publish') AND ( (wp_postmeta.meta_key = 'product_type' AND CAST(wp_postmeta.meta_value AS CHAR) = 'type_pre_ground')
OR  (mt1.meta_key = 'product_type2' AND CAST(mt1.meta_value AS CHAR) = 'type_pre_ground')
OR  (mt2.meta_key = 'product_type3' AND CAST(mt2.meta_value AS CHAR) = 'type_pre_ground')
OR  (mt3.meta_key = 'product_type4' AND CAST(mt3.meta_value AS CHAR) = 'type_pre_ground')
OR  (mt4.meta_key = 'product_type5' AND CAST(mt4.meta_value AS CHAR) = 'type_pre_ground')
OR  (mt5.meta_key = 'product_type6' AND CAST(mt5.meta_value AS CHAR) = 'type_pre_ground') ) GROUP BY wp_posts.ID ORDER BY wp_posts.post_date DESC

Cela ne s'est jamais terminé et a immobilisé le processeur à une charge très élevée. J'ai supprimé les deux derniers INNER JOIN (et les clauses WHERE correspondantes) et j'ai obtenu des résultats en 2 secondes :

SELECT   wp_posts.* FROM wp_posts  
INNER JOIN wp_term_relationships ON (wp_posts.ID = wp_term_relationships.object_id) 
INNER JOIN wp_postmeta ON (wp_posts.ID = wp_postmeta.post_id)
INNER JOIN wp_postmeta AS mt1 ON (wp_posts.ID = mt1.post_id)
INNER JOIN wp_postmeta AS mt2 ON (wp_posts.ID = mt2.post_id)
INNER JOIN wp_postmeta AS mt3 ON (wp_posts.ID = mt3.post_id) WHERE 1=1  AND ( wp_term_relationships.term_taxonomy_id IN (7,11,12,13,14,15) AND wp_posts.ID NOT IN (
                    SELECT object_id
                    FROM wp_term_relationships
                    WHERE term_taxonomy_id IN (10)
                ) ) AND wp_posts.post_type IN ('post', 'page', 'attachment', 'product', 'store_page') AND (wp_posts.post_status = 'publish') AND ( (wp_postmeta.meta_key = 'product_type' AND CAST(wp_postmeta.meta_value AS CHAR) = 'type_pre_ground')
OR  (mt1.meta_key = 'product_type2' AND CAST(mt1.meta_value AS CHAR) = 'type_pre_ground')
OR  (mt2.meta_key = 'product_type3' AND CAST(mt2.meta_value AS CHAR) = 'type_pre_ground')
OR  (mt3.meta_key = 'product_type4' AND CAST(mt3.meta_value AS CHAR) = 'type_pre_ground') ) GROUP BY wp_posts.ID ORDER BY wp_posts.post_date DESC

J'ai donc réalisé que la suppression des INNER JOINS était la clé pour accélérer la requête. J'ai réécrit la requête avec un seul wp_postmeta INNER JOIN et j'ai obtenu des résultats en quelques fractions de seconde :

SELECT   wp_posts.* FROM wp_posts  
INNER JOIN wp_term_relationships ON (wp_posts.ID = wp_term_relationships.object_id) 
INNER JOIN wp_postmeta ON (wp_posts.ID = wp_postmeta.post_id) WHERE 1=1  AND ( wp_term_relationships.term_taxonomy_id IN (7,11,12,13,14,15) AND wp_posts.ID NOT IN (
                    SELECT object_id
                    FROM wp_term_relationships
                    WHERE term_taxonomy_id IN (10)
                ) ) 
AND wp_posts.post_type IN ('post', 'page', 'attachment', 'product', 'store_page') 
AND (wp_posts.post_status = 'publish') 
AND ( (wp_postmeta.meta_key = 'product_type' AND CAST(wp_postmeta.meta_value AS CHAR) = 'type_pre_ground')
OR  (wp_postmeta.meta_key = 'product_type2' AND CAST(wp_postmeta.meta_value AS CHAR) = 'type_pre_ground')
OR  (wp_postmeta.meta_key = 'product_type3' AND CAST(wp_postmeta.meta_value AS CHAR) = 'type_pre_ground')
OR  (wp_postmeta.meta_key = 'product_type4' AND CAST(wp_postmeta.meta_value AS CHAR) = 'type_pre_ground')
OR  (wp_postmeta.meta_key = 'product_type5' AND CAST(wp_postmeta.meta_value AS CHAR) = 'type_pre_ground')
OR  (wp_postmeta.meta_key = 'product_type6' AND CAST(wp_postmeta.meta_value AS CHAR) = 'type_pre_ground') ) 
GROUP BY wp_posts.ID ORDER BY wp_posts.post_date DESC

Avec cette solution, j'ai créé un filtre WordPress pour convertir la requête postmeta au format plus rapide :

function custom_get_meta_sql( $meta_sql )
{
    global $wpdb;

    $posts_table = $wpdb->prefix . 'posts';
    $postmeta_table = $wpdb->prefix . 'postmeta';

    //use single INNER JOIN
    $meta_sql['join'] = " INNER JOIN {$postmeta_table} AS pmta ON ({$posts_table}.ID = pmta.post_id) ";

    //replace the mtNN aliases with wp_postmeta
    $where_clause = $meta_sql['where'];
    $where_clause = str_replace("{$postmeta_table}.", 'pmta.', $where_clause);
    $where_clause = preg_replace('/mt\d+\.meta_/i', 'pmta.meta_', $where_clause);

    $meta_sql['where'] = $where_clause;
    return $meta_sql;
}

add_filter( 'get_meta_sql', 'custom_get_meta_sql' );

Ce filtre doit se trouver dans le fichier functions.php de votre thème WordPress.

besoin d'aide pour optimiser wordpress meta_query