RE: CHAIN Versus SETLL and READ When Data Needed

You are correct Alan.  We are actually moving more to using embedded SQL
for file I/O and it is easier, usually.  However, there are still many
many programs doing it the traditional way as well as traditional
programmers who do not want to use SQL.

-----Original Message-----
From: rpg400-l-bounces@xxxxxxxxxxxx
[mailto:rpg400-l-bounces@xxxxxxxxxxxx] On Behalf Of Alan Campin
Sent: Wednesday, February 28, 2007 6:40 PM
To: rpg400-l@xxxxxxxxxxxx
Subject: RE: CHAIN Versus SETLL and READ When Data Needed

I know this is a question about which is more efficient but wouldn't the
answer be neither?

In other words, wouldn't this be much more efficient to use SQL
especially if you have multiple records? 

If it was only one record, it wouldn't matter but if you had many
possibilities, SQL would only bring in the records that matched your
condition. If all the matching records where being updated the same way
you could just issue an Update.

If not, wouldn't you issue a

EXEC SQL DECLARE UP_CUR CURSOR FOR
           SELECT PROJNO,
                  PRSTDATE,
                  PRENDATE
             FROM PROJECT
             FOR UPDATE OF PRSTDATE,
                           PRENDATE;

because you know every record is being updated?

Anyway, my two cents.