| order by RAND() . need a faster solution
|
nickCR
msg:4135659 | 2:26 am on May 20, 2010 (gmt 0) |
Hello All,
So here is a query i'm trying to run to get a randomized result from our db.
I've been looking for hours now on a solution and just don't seem to be finding what I need.
The difference is as such.
When I run the query without the order by rand() limit 1 I get a result in 0.0052
When I run the query with the above I get a result in 0.0517
It may seem like a decent result time but when the entire page i'm trying to load this query onto is loading in 0.03 that is adding quite a bit of extra latency (0.07 / 0.08).
Here is the query:
SELECT t3.idS, sum(t1.value - t1.price) as savings
FROM table1 t1
INNER JOIN table2 t2 ON t1.id = t2.idC
INNER JOIN table3 t3 ON t3.idC = t1.id
WHERE t3.idS NOT IN (5653,21740)
GROUP BY t3.id
HAVING sum(t1.value - t1.price) > 100
ORDER BY RAND()
LIMIT 1
Any ideas what I can do to avoid the extra result time?
table1 = 598
table2 = 598
table3 = 10285
BTW if I remove table2 from query it takes almost double the time.. I have no idea why since it's not even being used in any of the values.
I would truely appreciate any help.
Thanks in advance,
Nick
|
zergtant
msg:4135683 | 3:20 am on May 20, 2010 (gmt 0) |
try this one use min and max with the Primary Key id
SELECT *
FROM `table` AS t1 JOIN (SELECT ROUND(RAND() * ((SELECT MAX(id) FROM `table`)-(SELECT MIN(id) FROM `table`))+(SELECT MIN(id) FROM `table`)) AS id) AS t2
WHERE t1.id >= t2.id
ORDER BY t1.id LIMIT 1;
and btw i say you use multiply join and GROUP BY, may be you can create index to make select faster
|
nickCR
msg:4136012 | 4:57 pm on May 20, 2010 (gmt 0) |
Thanks for your suggestion.
However after spending hours and hour last night trying to find a faster way to do this I think it actually makes more sense to run a process once a day after we run our financial batch to re-calculate the "savings" and add it to the members table and then do a simple select following some of the solutions that already exist out there.
|
serutan
msg:4139775 | 7:17 am on May 26, 2010 (gmt 0) |
How about this: generate a random string which you pass to your procedure as a parameter, and order your results by this string bitwise ANDed with some string field in your data. Something tells me that would run faster than having mysql generate a random number for every row, but I really dunno.
The main question that springs to mind is, I know efficiency is a good thing but will eliminating a few hundredths of a second of processing time in this particular process pay you back for your efforts?
|
|
|
