1:11 AM
To: u2-users@listserver.u2ug.org
Subject: RE: [U2] [UD] Indexing 'Best Practices'
Hi David,
Don't know if this is possible or not but couldn't you write your COMPLETE
records to a .HIST file, keeping the rest of the data in a .CURR file. Then
have a PART file definition that bind
rd
Senior Programmer/Analyst
Harbor Wholesale Grocery
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of David Wolverton
Sent: Wednesday, August 02, 2006 6:53 AM
To: u2-users@listserver.u2ug.org
Subject: RE: [U2] [UD] Indexing 'Best Practices'
OK
"Kate Stanton" <[EMAIL PROTECTED]> wrote on 2 Aug 2006 18:11:11 +1200
> I would go for the simplest solution to operate and maintain,
> almost to hell with machine efficiency. You (or anyone else
> trying to follow you) are probably the most critical and
> expensive resource.
I must disagree wit
At 20:59 -0500 2006/08/01, David Wolverton wrote:
So we have tens, and hundreds of records (and have to think about thousands)
in the different status 'buckets' and can grab those instantly via a SELECT
- and that is great and works like a champ -- except that we are also
carrying an index for CO
OK - sold! Virtual field to move COMPLETE to NULL is the answer... As I was
driving back to the hotel after doing that post last night, I got to
thinking that the Virtual Field computation would strictly be CPU/Memory,
which means I can do A LOT of that for the cost of a single disk access.
And wi
]>
To:
Sent: Wednesday, August 02, 2006 1:59 PM
Subject: [U2] [UD] Indexing 'Best Practices'
I have a scenario to describe and would like comments as to what others see
as a 'Best Practice' for Indexing. We're on UniData, currentish Windows
version, and have indexes in pl
>From: David Wolverton
>Would we be better off to create a "Virtual" field definition to
index that carries the 'real' code into the index, unless it is
COMPLETE, in which case we would make the virtual field pass a NULL
and then index this new virtual field using NoNulls?
Assuming COMPLETE can't
I have a scenario to describe and would like comments as to what others see
as a 'Best Practice' for Indexing. We're on UniData, currentish Windows
version, and have indexes in place on files that are starting to grow at
larger client sites and are heavily modified.
Here is a common index issue we