Hello Peter,
thanks for your reply, however my question was less related to RAND()
but more about the weird situation that the results are more or less
fine with the two queries, however different when they are combined
into one query with a subquery.
Thanks
--- Peter Brawley <[EMAIL PROTECTED]
Neroux,
RAND() gives a roughly rectangular distribution, not a normal
distribution, so samples of fewer than 100 or so values from it are
likely to show large but statistically _insignificant_ differences. In
larger samples, I don't see any tendency of RAND() to produce more
values near its
Hello,
I am having a table with an integer column which I use for probability
calculations (the higher a value the more often it is chosen).
Now I am having the following query, which should actually incorporate
these probabilities, however it seems to prefer values from the middle
range
SELE
