Re: Which of the SYSIBM tables/views show the row count for

[Joe Pluta <joepluta@xxxxxxxxxxxxxxxxx>]

rob@xxxxxxxxx wrote:
> So the UNIQUE was a good one in this case.  I think it would be fair to 
> say that for files of this genre, in which there should be no duplicate 
> keys, then it should be a good one in all of that genre.  Unless you can 
> come up with a good business reason why not, and can give a sample 
> thereof.
>   
I thought I should do the runs myself, and I'm getting significantly 
different results, Rob.

With no key at all, I get the 330ms number.  With a key but without 
UNIQUE, my average jumps to 1450ms.  Add UNIQUE to the file, and it 
leaps to 9300+.  You might want to check your results.  I did it by 
actually performing a CHGPF on the file each time and clearing it before 
each run.

And this is for only a single key, but at the same time no data to speak 
of.  I suspect the numbers change based on the size of the record and 
the size of the key.  However, if my numbers are correct, then a 7-to-1 
jump in overhead is nothing to sneeze at, and it requires a darned good 
business reason to add the keyword.

Joe