Question

I am trying to get a random row in MySQL. Here is my approach:

SELECT * 
FROM users 
WHERE id = 
 (SELECT floor((max(id) - min(id) + 1) * rand()) + min(id)
  FROM users);

This can return an empty set when there are gaps in the ids. I am fine with that. However,

1. I don’t have gaps and I still get an empty set rather frequently.
2. More disturbing: I will get two or more results every once in a while.

What is the reason for this strange behavior? How, do I get it right?

Note:

• the table is quite big (10^6 … 10^7); I can’t use any of the obvious, and known, solutions based on random ordering or even based on using count().
• I’am stuck with MySQL, I don’t care how it’s done on the client side ruby/PHP/whatever.
• I tried to cast the floating types, but it did not help.
• The corresponding PostgreSQL variant (random instead of rand, and some type casts) works fine.

Answer 1

This avoids repeated evaluation of random() during the aggregation and should be faster, too, as min(id) and random() are only evaluated once.

SELECT u.* 
FROM   users u
      ,(SELECT min(id) AS min_id, max(id) AS max_id FROM users) x
WHERE  u.id > (floor((x.max_id - x.min_id + 1) * rand()) + x.min_id)
LIMIT  1;