Potentially extraneous homepage links

goal: find extraneous external links like in "[http://example.com/page foo] on [http://example.com/ example site]" - without parsing dumps, but with reasonably few false positives

SQL

USE barwiki_p;
SET @language = 'bar';
SELECT CONCAT('https://', @language','.wikipedia.org/wiki/',page_title) AS page_url, 
rooturl,
SUM(IF( url=rooturl, 1, 0) ) AS rootlinks,
SUM(1) AS alllinks
FROM (
  SELECT
    page_title,
    el_to AS url,
    CONCAT(SUBSTRING_INDEX(el_to, '/', 3),'/') AS rooturl
  FROM externallinks, page
  WHERE el_from = page_id
  AND page_namespace = 0) AS pagelinks
GROUP BY page_title, rooturl
HAVING rootlinks > 0 AND rootlinks < alllinks
  AND LOCATE(REPLACE(LOWER(CONVERT(page_title USING utf8)), '_', ''), rooturl) = 0
LIMIT 200

SQL

Checking query status...