Re: Replication, GUIDs and PKs

Tech-Archive recommends: Repair Windows Errors & Optimize Windows Performance

I did a test and guid tables indexes to be 4 times as fragmented as bigint
cols.

--
Hilary Cotter
Director of Text Mining and Database Strategy
RelevantNOISE.Com - Dedicated to mining blogs for business intelligence.

This posting is my own and doesn't necessarily represent RelevantNoise's
positions, strategies or opinions.

Looking for a SQL Server replication book?
http://www.nwsu.com/0974973602.html

Looking for a FAQ on Indexing Services/SQL FTS
http://www.indexserverfaq.com



"Michael MacGregor" <nospam@xxxxxxxxxx> wrote in message
news:uSv8T$O3GHA.1464@xxxxxxxxxxxxxxxxxxxxxxx
To be honest I don't know at this point. I am in the investigation and
analysis stage.

I am trying to pull together a whole bunch of information before I embark
on any monitoring of what is actually going on under the cover. However,
as far as fill factor is concerned, it's set to the default for all
indexes.

Changing the fill factor though can impact read performance, and there are
already some schema design issues that are affecting read performance that
I definitely do not want to become worse.

Also, from a lot of what I have come across so far with regard to the use
of GUIDs as PKs, it is strongly advised not to create a cluster on those
PKs, which makes sense to me. The addition of the sequential GUID in 2005
would seem to have been introduced in order to address the concerns of
using non-sequential GUIDs, however, that particular option is of no use
as the subscriber databases are all 2000, and besides it isn't SQL that is
generating the GUID anyway, but the application.

There are a number of issues with the design, implementation and
maintenance of the databases here that I was brought in to address, but it
would seem that they are related issues and if I change one thing it will
impact another.

Michael MacGregor
Database Architect



.

Relevant Pages

  • Re: GUID default value
    ... > SQL Server and MS Access databases. ... We have default values for the GUID column in Oracle DB ... Autonumber GUID (called a "Replication ID"). ... I don't use Jet DDL since it's not as full-featured as DAO, ...
    (microsoft.public.access.replication)
  • Re: GUID default value
    ... > We are developing a complex system, which will run on Oracle, MS SQL ... > All the databases are to be replicated. ... > GUID column, which is filled once a new record is inserted. ... Jet has a GUID data type. ...
    (microsoft.public.access.tablesdbdesign)
  • Re: GUID on Linux
    ... >> variations. ... starts with the MAC because that is unique to the network card (across ... Ok, at this point it is a Globally Unique IDentifier, or guid. ... objects in databases which are then moved around in other databases. ...
    (comp.os.linux.development.apps)
  • Re: DB2, MsSQL, and MS Access, How to maintain 3 data sources
    ... Don't seem to be a good idea to use GUID. ... He will have to change the data structure for all his tables and databases, ... They are guaranteed unique world-wide and make the most sense ... >>> unique customer ID's? ...
    (microsoft.public.vb.database.ado)
  • Re: SQL 2005 Distribution DB not visable
    ... Director of Text Mining and Database Strategy ... RelevantNOISE.Com - Dedicated to mining blogs for business intelligence. ... Looking for a SQL Server replication book? ... User Databases. ...
    (microsoft.public.sqlserver.replication)