Re: native P.P. in V5R4???

Walden,

You are correct about some cases where "native" I/O has been faster
than SQL.  The same argument was made in the 1970s & 80s  between SQL
and DB2 and IMS "native" access on the mainframe.  But, that platform
realized it needed to compete in the RDBMS world and the programmers
moved to the modern world.

You're also correct about looking at the problem from a different point
of view.   I'll add..., with a different design architecture, 
methodology and perspective.  

It is true that SQL is not the end-all, be-all, but is the standard
language (OK, Joe, data access method if you're a nit picker) if you're
a serious contender in the RDBMS world.    The iSeries's serious
competition is Microsoft and UNIX-flavors running internationally known 
RDBM'S like SQL Server or ORACLE.   The don't use READ, WRITE and CHAIN.
  They use SQL with HLL.  The folks buying the platforms, don't even
consider the iSeries, in many cases, because it doesn't have an
internationally recognized RDBMS environment because still touts RPG (to
those folks that means Rocket Propelled Grenade) and they know most of
the apps in RPG using the legacy record access methods. 

Let's work together on moving this platform forward and not staying in
the past.

Take care,

Dave 

>>> WaldenL@xxxxxxxxxxxxxxx 2/23/2006 15:28:26 >>>
>SQL works great for result sets but I would have a hard time 
>believing that single record SQL access is as fast as RPG 
>native DB access.  Maybe things have changed and my brain data is old
though. 

I'm a SQL lover, but I think it's been shown over and over that there
are cases where "native" IO blows the doors of SQL. 

HOWEVER, you must also take a step back and look at the problem from a
different point of view. I've seen cases where SQL was substituted for
a
CHAIN or a SETLL/READE and it was "shown" that the result was slower
than the native IO. But the real question is, would you have even
needed
that CHAIN if you'd looked at the problem from a SQL point of view and
used a join? Or would you need to read all those records if you looked
at the problem from a SQL point of view and used a "sum(fld1) group by
fld2" SQL statement.

SQL is damn fast in set oriented tasks, the real trick is to see if
you
can rework the initial question to be set oriented instead of trying
to
make SQL solve a problem that's not set oriented. SQL really doesn't
mind 10-way joins w/subselects, and sometimes that's how you need to
look at the problem.

-Walden

------------
Walden H Leverich III
Tech Software
(516) 627-3800 x3051
WaldenL@xxxxxxxxxxxxxxx 
http://www.TechSoftInc.com 

Quiquid latine dictum sit altum viditur.
(Whatever is said in Latin seems profound.)