HP3000-L Archives

September 2000, Week 3

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:
"Glenn A. Mitchell" <[log in to unmask]>
Reply To:
Glenn A. Mitchell
Date:
Mon, 18 Sep 2000 10:09:03 -0600
Content-Type:
text/plain
Parts/Attachments:
text/plain (50 lines)
We've used an X8 key field for our MPI (Master Patient Index!!) since the
days when we were on a Series 48.  Our numbering scheme goes from 0000001 to
9999999 left justified in the X8 field.  The reasoning for the key structure
goes back to the systems we were replacing and also reflects the sparse
nature of the keys (numbers beginning with 0, 1, 2, or 3 reflect numbers
assigned prior to our "new" system; those beginning with 4 were assigned
with our current system; those with 9 at the beginning were assigned during
system outages).  Using an integer key just wasn't an option for us.

Performance with this scheme is certainly good.  Howmessy doesn't show
exceptional secondaries and response times are very good on our 957 with
about 700,000 numbers assigned.

--
Glenn A. Mitchell  Mailto:[log in to unmask]
3GM Associates, Inc
Portland, ME  04102
207-772-9370
"Ron Burnett" wrote in message news:39c54d4a$1_1@skycache-news.fidnet.com...
> Greetings everyone,
>
> A question for the IMAGE gurus:
>
> For many years we have had a PMI (Patient Master Index) with a
> primary key of X6 wihch contains an all numeric identifier in the
> range of 100000 to (now) about 940000.  This has served us very
> well, with pretty good retrieval performance, even with a severely
> overloaded system.
>
> However, we estimate that we are going to hit the 1000000 mark
> in issued numbers within the next 14 to 18 months.  Time to think
> about expanding that X6 to an X8.
>
> My concern is what it will do to storage efficiency and retrieval
performance
> if we just take all those existing key values and change them to X8 with
> right justify and leading-zero-fill.  I don't like the sort problems if we
> left
> justify and pad with spaces (although, to be honest, a numeric sort is
> very infrequently done).  Right-justify with leading SPACE fill is a
> possibility,
> if it gains any advantage.
>
> Any insight will be highly appreciated.
>
> Cheers from the land of 27th Olympiad,
> Ron Burnett
> [log in to unmask]
>

ATOM RSS1 RSS2