Re: Another use of SQL RAND

Chuck,

Well, Shalom's posting was good, but this is brilliant!

There's something that I don't understand. The following lines:

select int(rnum+1) , decimal(rand(), 10, 10)
from sample
where rnum<100

It seems to retrieve recursively only the *last* record of the table
"sample" on every recursion. In a recursive query, is that always the rule?


"CRPence" <crp@xxxxxxxxxxxxxxxxxxxx> wrote in message
news:mailman.2714.1203003999.1203.midrange-l@xxxxxxxxxxxxxxx

Why deal with an identity, or especially inserting n-times, just
generate the values with a recursive query for the number of rows
required:

<code>
CREATE TABLE shuffle (nid , rnd ) as
( with sample (rnum , ranv) as
( select int(1), rand()
from qsys2.qsqptabl
union all
select int(rnum+1) , decimal(rand(), 10, 10)
from sample
where rnum<100
) select rnum,ranv from sample
) with data
</code>

Regards, Chuck
--