Indexed Subfiles and QTP

[Aikens, Curtis]

Bear in mind however that sequential access, on an indexed file, still means
reading the file in primary key order (using RMS anyway).  This by it's
nature incurs overhead not associated with reading a true sequential file.

Of course, your key definition comes into play as well - lots of duplicates
will cause lots of overhead too.

We once had an A/R system which took just over a day and a half to process a
day's worth of transactions (uh hello?  Anybody heard of volume testing
BEFORE implementation?).   It was resolved by improving the file key design
and declaring some null keys.

I like (and am interested) Allison's test as well.  It would be interesting
to see if an indexed subfile has better or worse performance than an indexed
file created through the dictionary.  I'd still bet that a true sequential
file will blow away your best, most optimized index file, even on a pure
sequential read...

But I've been wrong (once) before!  =8^)

Curtis Aikens
Information Technologies
Dairyworld Foods 
(780) 486-8442

 -----Original Message-----
From: 	Merol Newman [mailto:merol.newman@ramesys.com] 
Sent:	Thursday, January 11, 2001 2:38 PM
Subject:	RE: Indexed Subfiles and QTP

Hello Robert
[C.Aikens] <snip>

I've just seen some other replies - it occurs to me that default access IS
sequential in VMS or Unix C-ISAM, but quite possibly not in MPE.

regards
Merol

= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Subscribe: "subscribe powerh-l" in message body to majordomo@lists.swau.edu
Unsubscribe: "unsubscribe powerh-l" in message to majordomo@lists.swau.edu
This list is closed, thus to post to the list, you must be a subscriber.