QTP Performance - Clarification

[John Webster]

I'm coming into this discussion a little late, but I once wrote a 4-line QTP
program on the HP3000 that took 4 days to run... one line per day?

>From your description of the data here's what I suspect the problem MIGHT be:
the Pareto principle at work. How many nationalities are involved? Are they
evenly spread, or are 90% the same. IF you are keyed on nationality, you will
have a huge number of duplicate keys; each time you change one, KSAM has to
rebalance its B-tree. My suggestion would be to run a Quiz report over the data
counting the number of instances of each nationality. Pick those that give you
90% of the data and run another Quiz to create a subfile (unindexed) of the
90%, then run QTP to update the subfile, using item statements rather than lookups
(with a CASE option for more efficiency). Then recreate a new KSAM file from
the subfile. Then deal with the remaining records in the way you have been processing
so far...

I think the problem is in KSAM, not in QTP.
Cheers
John W 
----------------------------------
John Webster
Wellington, New Zealand
One of the World Wide Websters
------------------------------