Re: Oracle to PGSQL -- need help

On Thu, 25 Jul 2002 15:17:08 -0600
Randall Barber <rdb55@email.byu.edu> wrote:


> I greatly appreciate your time and help--I have not yet tried to implement
> what you gave me, but here is some additional explanation:
>
> ABS(MOD(DBMS_RANDOM.RANDOM, 100))
>
> This generates a random number between between +/- 2^32.   I use the MOD to
> keep the number with a range (0-100) and an ABS to make it positive
> regardless.  The idea is to pick a random row out of information that shares
> a common value.   In your example, you list everything in the table, but
> without much changing I think I can make it so that the result brings back a
> single row instead of multiple.  Just needs a where clause in one of the
> SELECTs :).


Thank you for your explanation. I were roughly able to figure out the whole
process you thought. But there are some columns I don't know; actually
their names don't appear in either of this mail and prior mail. Could you tell
me the answer to the following questions ?


 1. Does "a common value" mean one of the values that Q column has?
 2. Does "a range (0-100)" mean the fact that the range is always 0 to 100 ?
    Doesn't the range depend on the number of the rows that share a selected
     value ?



>> Given an insert request, I want the trigger to find out which # is used on
>> the left, and then randomly select one of the resultant rows and insert the
>> randomly picked row's PKEY value along with the original insertion info.
>> into a third table.

3. What table is "a request" inserted into ?
4. What column is "PKEY" created on ?



>> Q, JOBID
>> --------
>> 1,2
>> 1,5
>> 1,200
>> 1,4356
>> 1,483

4. Is this a second table ?



Regards,
Masaru Sugawara