The question is not really clear; what is the problem;
is it slow due to seq scans??

try to get a explain and analyze it or when the program runs
or
do a onstat -g tpf <threadid> and check for seqscans; or look in
sysmaster:sysptprof for these tables

assuming it's seq scans:
Have seen issues like this. not only 4gl but also in spl.

question: where is OPTCOMPIND set to;
if 2 then use 0!!!
Have also seen this boosting performance with factors!!

you may getaway without statistics on tables. !!

Superboer.