Im probably not making myself clear.
Ultimately, my goal is to emulate mysql's ->

for other databases that do not support RAND(). So using variations of php's rand(), wouldn't make sense,
as it only picks one value out of a range of values, that are numerical in nature.

But here is an idea, that Im thinking about, but haven't gotten to the code as yet.
1. In the tables that I want random values from, create a "rand" column, that contains incremental numerical values for each row.
2. select the the maximum number from the rand column.
3. Assume that there are no gaps between 0 and this max number.
4. Create an array of numbers with values between 0 and max number.
5. Use array_rand() to "randomly" choose (x) values from the array created in step 4. (I may choose maybe
a 1 or 2 values more than whats required, just in case of gaps between 0 and max number in step 3).
6. Use these randomly choosen values to select from the database as random rows.

So hopefully its a bit clearer what Im striving for.
To me the idea above would work, but it hinges on if that rand column, doesn't have gaps.
If you see room for improvement, or have another idea, or if Im talking gibberish then by all means.


Doug Thompson wrote:

An incredible interpretation of

If called without the optional min, max arguments rand() returns a pseudo-random value between 0 and RAND_MAX. If you want a random number between 5 and 15 (inclusive), for example, use rand (5, 15). </quote>


PHP Database Mailing List (
To unsubscribe, visit:

Reply via email to