Re: Need help in reducing processing time...

At 21:16 06/04/2002, Ray Nainy wrote:
>I'm trying to explore the Best options to return the Result Set to Client
>Server application. Result Set contains data from different files and with
>different sort orders. We are not allowed to use the Join Logicals. So, is
>there any better way to improve the processing time other than using arrays?
>My goal is to minimize the processing time. Any suggestions are greatly
>appreciated.

Assuming that you are using ODBC or JDBC, you can use an SQL select and
return it as a result set cursor. Depending on the query, I have had to
resort to building a temporary result table in QTEMP and returning the
result cursor from that in order to allow efficient retrieval of ancillary
data and result ordering, but even then, it can still be reasonably quick.
Run the queries using visual explain (or just in debug if you prefer) in
order to identify any indexes that the optimizer wants. Providing them can
make a major difference. If only a few rows are returned, it can help to
use an OPTIMIZE clause so the optimizer knows how many rows to expect. If
you can specify FOR FETCH ONLY AND WITH NC or NONE for commitment, that can
help too by letting the system know it can block and doesn't need to lock rows.


Pete Hall
pbhall@ameritech.net
http://www.ameritech.net/users/pbhall/index.html