Years ago. Don't want to get into a big deal with this.
IBM sent a representative to Common to talk about performance issues. He
indicated in the first release of the System 38, IBM did a read of the buffer
for READE but ran into problems so they changed the behavior of READE to flush
the buffers each time a READE was done, like a CHAIN. IBM called this a Logical
I/O and claimed nothing took up more resources on an AS/400 than issuing a
Logical I/O. In contrast, they indicated that a READ was implemented by RPG and
read from the buffer looking for the next record. A process they indicated was
a lot more efficient.
I am just passing along what I was told. It is possible that IBM has changed
the behavior of READE but I doubt it. Maybe somebody from IBM will speak to the
issue.
I like the code because it is simple and easy. The SETLL sets the position and
there is only one read. I use the DoU because I always go into the loop no
matter what the result of the SETLL. SETLL just sets the position.
-----Original Message-----
From: Paul Morgan [mailto:pmorgan@xxxxxxxxxxxxxx]
Sent: Tuesday, March 29, 2005 3:03 PM
To: rpg400-l@xxxxxxxxxxxx
Subject: Re: Chain doesn't set %EOF
Alan,
Had a prior thread on DoW vs DoU read loops.
Who the heck was the IBMer who said never to use ReadE? How long ago was
the Common conference? Possible old performance problems with ReadE?
I doubt Barbara would agree with this recommendation to never use ReadE.
Paul
--
Paul Morgan
Senior Programmer Analyst - Retail
J. Jill Group
100 Birch Pond Drive, PO Box 2009
Tilton, NH 03276-2009
Phone: (603) 266-2117
Fax: (603) 266-2333
"Alan Campin" wrote
> My recommendation.
>
> SetLL(SomeKey) File;
> DoU 0 = 1;
> Read (File);
> If %Eof(File) or
> SomeKey <> KeyField;
> Leave;
> EndIf;
> // Do something
> EndDo;
>
> Was told by IBM to never use a READE, always use a READ. From a Common
conference years ago.
