HP3000-L Archives

December 2001, Week 4

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Sletten Kenneth W KPWA <[log in to unmask]>
Reply To:
Sletten Kenneth W KPWA <[log in to unmask]>
Date:
Wed, 26 Dec 2001 13:21:06 -0800
Content-Type:
text/plain
Parts/Attachments:
text/plain (40 lines)
Dave sez:

> My personal belief is that if your average chain lengths get
> to be very long (7 or more, say...) then use an IMSAM key if
> you need sorted retrieval.

This is one of those "it depends" situations;  and it probably depends
more than most cases.  Key considerations:

(1)   How often will entries on the sorted chain have to be re-sorted
during normal operations ??..  If answer is "a lot", then certainly you
have to be careful about keeping chain lengths within reasonable
bounds....   But if answer is "almost never" or at least "rarely", then
on balance the (sometimes) huge benefit of keeping data always
available in sorted order for reporting purposes may be compelling;
even if the chains get fairly long.

(2)   Will new entries on the sorted chain mostly or always be added
at the end of the chain ??..  If not, the same "be careful" as in (1)
above applies....  But if answer is "always" you probably don't care
too much how long the sorted chain gets, since TurboIMAGE
(currently) starts looking at end-of-chain for the "right" slot for a
new entry....


We have a few sorted chains in our database where the global
average chain length (GACL) is between 20 and 60, and end-users
do not have any perceptible performance problems......  because the
data making up those sorted chains is essentially static...  Just did a
quick scan, and noticed we have ONE sorted chain where the GACL
is over 700.....   crazy, you think ??....  well..:  In this case the sort
field is DATE, so new data is always added at end of the chain.....
and existing data on the chain amounts to an historical archive that
normally never changes....

Ken Sletten

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *

ATOM RSS1 RSS2