Re: Running SQL commands from CL (or elsewhere)

Dave,

As much as I prefer to do stuff in RPGLE over CL there are a few 
applications (like backups) where I'll stick with CL.  When doing so, 
sometimes having a Q&D method to execute CL comes in handy.  Several 
people use an old command from a magazine called RUNSQL.  Basically it is 
a QMQRY that is nothing but parameters and you pass it your entire 
statement.

STRSQL is not very conducive to a batch job based on CL.  And an 
interactive one that stops and presents a user with the STRSQL screen 
would go over like a turd in the punchbowl.

Building some stock queries for use with STRQMQRY is a possibility. 
However, instead of building a series of stock queries, why not use the 
qshell technique?  Granted, there might be a performance hit because of 
starting the environment, but is it a significant performance hit?  Most 
SQL I see imbedded into CL is a set, versus row, transaction, that is 
executed only once in the program.

To me, it's just another tool in the toolbox.  My only remaining question 
is "What about error trapping?".  But honestly, most of the sql I imbed 
into CL with RUNSQL I do no error trapping.  A lot of "assume" is done. 
And y'all know where that can lead.

Rob Berendt