akinonyx

Problem: I'm trying to build a junction table for a many to many relationship. Table A contains a title string that contains in it names that are stored in Table B. Both tables have nearly a million rows. Bad solution: The method I'm using right now is using PHP I fetch and store one table into an associative array. I then do a mysql_fetch loop on the second table and run this code to populate an array that is used to build the multi-insert SQL. $title = preg_replace ($names, $keys, $title); if (preg_match_all ('/(\d+)/', $title, $matches)) foreach ($matches[1] as $m) $inserts[$row['id'].'-'.$m] = 1; This works, but I'm looking for possibly a less cpu/memory intensive method. My remote server can't handle it, and we're talking hours and hours on a local server. It seems the bottleneck is with the preg_replace. I've also tried str_replace, which is only slightly faster. So, a method where I wouldn't have to pass around such large arrays, or an entirely different approach is what I'm looking for. Any ideas?