 |
 |
Archives of the TeradataForum
Message Posted: Fri, 28 Jun 2002 @ 14:59:39 GMT
| Subj: | | Re: Uniform Distributions from RANDOM number function |
| |
| From: | | Cornelius Meade |
As you point out even an ideal random function could be expected to at times generate clusters of values. The only way I have
ever been able to do something similar to what you are doing (which was not related to Teradata unfortunately) while minimizing the
type of problem you are having was when I was able to control the seed value that was used by the random number generator. In my
case, I developed a series of seed values which I had tested in conjunction with each other through the random number generator
function I was using ensure that they would generate as even as possible a distribution of values. It sounds like maybe the
algorithm that is selecting the seed value in your RANDOM function, or the value generator algorithm itself is not as random as
perhaps you need it to be.
| |
