Okay, here is something I found out, which is working for my use case and might be helpful as a starting point for someone in the future. There might be better and more elegant ways to do this, but at least it does the trick in some cases.
Background: I use a post_relation field (among other) inside a repeater field and want to query all posts, which do not have a relation to one or more given post ids. As an example: I query posts of a post_type ‘recipe’ and want to get all posts (=recipes) which dont have a relation to some ingredients (‘ingredient’ is a post_type as well). The repeater field is named ‘ingredients’ and the relation field inside the repeater is called ‘ingredient’.
I build my wp_query step by step: standard query_args, meta_query_args and so on. Then I have an array of unwanted recipes. If this array is set (has values) I build a subquery and simply attach it at the end of the where clause (maybe not the best spot in regards to performance?). All of this happens in a closure (also, maybe not the best way to do this?).
if ( isset( $without_ingredients ) ) {
add_filter ('posts_where', function($where) use ($without_ingredients) {
global $wpdb;
$sql_in_no_repeater_extension = " AND $wpdb->prefix" . "posts.ID NOT IN (SELECT post_id FROM $wpdb->prefix" . "postmeta WHERE meta_key LIKE 'ingredients_%_ingredient' AND meta_value in (" . implode (',', $without_ingredients). "))";
return $where . $sql_in_no_repeater_extension;
});
}
What the simple subquery does, is getting the post id of all posts, having one of the uwanted post (ingredient) ids inside their ‘ingredients’-repeater subfields. Then these ids get removed from the original query.
So, maybe this helps someone sometime. If anyone has a hint how to do this better, I’d realy appreciate to read them.
