Re: MySQL rand() is slow on large datasets

[Jeff Fillmore]

I found that Jay solution gave a result set that were to close together. For example I would get rows from 7100 - 7150. That was not random enough for me. So here is what I did to get more random of rows with out killing the speed of Jay's query.

SELECT * FROM Table T JOIN (SELECT MAX(ID) AS ID FROM Table) AS x ON T.ID >= FLOOR(x.ID*RAND()) LIMIT 10;



Edited 1 time(s). Last edit at 05/21/2008 02:39PM by Jeff Fillmore.