Re: BLOCK(*YES) questions

["Booth Martin" <Booth@xxxxxxxxxxxx>]

I was taught that blocks were brought into memory  as chunks of disk storage
and would be some multiple of the disk's sector size.  The chunk has
whatever it has, which may be in key order, or mostly keyed order.  There is
no assurance that the block has the records in any special order, but the
chances are that the records will be in RRN order.  If the file keys are
incremented with an ascending key then the key and the RRN order may be
similar, in which case the block would also have the records in keyed order.
 A reorganized file will most likely be in RRN and key order.  Using deleted
records over again changes it all though.
 
 
 
---------------------------------------------------------
Booth Martin   http://www.MartinVT.com
Booth@xxxxxxxxxxxx
---------------------------------------------------------
 
-------Original Message-------
 
From: RPG programming on the AS400 / iSeries
Date: 02/27/04 11:07:35
To: rpg400-l@xxxxxxxxxxxx
Subject: BLOCK(*YES) questions
 
BLOCK(*YES) questions:
 
     FCMMASTER  IF   E           K DISK    block(*yes)
 
Does BLOCK(*YES) bring in a block of records in sequential order from the
file, or in keyed order?
 
RRN  Key
  1   AA
  2   ZZ
  3   EG
  4   BH
  5   WY
  6   WZ
  7   AB
 
In this scenario, will the block of records be in RRN sequence or in keyed
sequence?
 
RPG documentation is not, IMO, not very helpful in indicating how blocking
occurs.
 
The archives has a post that appears to support a block "by keyed records":
 
http://archive.midrange.com/rpg400-l/200212/msg00433.html
> 2. Blocking is done is in sequence that the records are being accessed:
> keyed or arrival.  NBRRCDS is always done in arrival sequence.  That's why
> the warning.
 
I have found several posts that imply that blocking is by physical sequence.
But nothing that says this.
 
Is there an IBM reference to this?
 
TIA,
Dan