Sample random revisions from big namespaces

SQL

# http://jan.kneschke.de/projects/mysql/order-by-rand/
# https://stackoverflow.com/questions/249301/simple-random-samples-from-a-sql-database
# https://mariadb.com/kb/en/library/data-sampling-techniques-for-efficiently-finding-a-random-row/#case-consecutive-auto_increment-without-gaps-10-rows
 
### All non-deleted revisions in the article namespace
#SELECT revision.rev_id, page.* FROM revision
#JOIN page on revision.rev_page = page.page_id
#JOIN (SELECT CEIL(RAND() * (SELECT MAX(rev_id) FROM revision)) AS rev_id) AS random
#WHERE revision.rev_id = random.rev_id
#AND rev_deleted = 0 AND page_namespace = 0;

 -- more than 10 (to compensate for dups)
SELECT @min := MIN(rev_id), @max := MAX(rev_id) FROM revision;
SELECT DISTINCT revision.rev_id, page.* FROM revision
JOIN page on revision.rev_page = page.page_id
JOIN (SELECT FLOOR(@min + (@max - @min + 1) * RAND()) AS rev_id FROM revision LIMIT 11) AS random USING (rev_id)
WHERE rev_deleted = 0 AND page_namespace = 0
LIMIT 10;

### All non-deleted revisions in the article-talk namespace
#SELECT * FROM revision JOIN page on revision.page_id = page.page_id WHERE rev_deleted = 0 AND page_namespace = 1 ORDER BY RAND() LIMIT 10000;

### All non-deleted revisions in the user namespace
#SELECT * FROM revision JOIN page on revision.page_id = page.page_id WHERE rev_deleted = 0 AND page_namespace = 2 ORDER BY RAND() LIMIT 10000;

### All non-deleted revisions in the user-talk namespace
#SELECT * FROM revision JOIN page on revision.page_id = page.page_id WHERE rev_deleted = 0 AND page_namespace = 3 ORDER BY RAND() LIMIT 10000;
SQL

Checking query status...
