Home | Contact Us | FAQ | Search & Site Map | Link to Us
Sign In | Join | Other 45 Sites in Network
Home
Discussion Groups
DB Engine
SQL ServerMSDESQL Server CE
Services
Analysis (Data Mining)Analysis (OLAP)DTSIntegration ServicesNotification ServicesReporting Services
Programming
CLRConnectivitySQLXML
Other Technologies
ClusteringEnglish QueryFull-Text SearchReplicationService Broker
General
Data WarehousingPerformanceSecuritySetupSQL Server ToolsOther SQL Server Topics
DirectoryUser Groups
Related Topics
MS AccessOther DB ProductsMS Server Products.NET DevelopmentVB DevelopmentJava DevelopmentMore Topics ...

SQL Server Forum / General / Data Warehousing / December 2004

Tip: Looking for answers? Try searching our database.

How to handle heirarchies in dimension attributes (Fairly green design question)

Thread view: 
Enable EMail Alerts  Start New Thread
Thread rating: 
Louis Davidson - 13 Dec 2004 19:27 GMT
For example, say I have a purchase fact table.  Then I have a customer
dimension.  In the customer dimension, we have city, state, name, etc, and
then we have a group level that needs to have a heirarchy exposed.  For
example:

Super
       Super Level 1
       Super Level 2
Good
       Good Level 1
       Good Level 2

Obviously this is not what we are really using, but it is the same sort of
thing.

How to best do this?  Using a related table?

Also, if there is a good way to search for this kind of discussion on
groups.google.com, advice there would be graciously accepted :)

Signature

----------------------------------------------------------------------------
Louis Davidson - drsql@hotmail.com
SQL Server MVP

Compass Technology Management - www.compass.net
Pro SQL Server 2000 Database Design -
http://www.apress.com/book/bookDisplay.html?bID=266
Note: Please reply to the newsgroups only unless you are interested in
consulting services.  All other replies may be ignored :)

Adam Machanic - 13 Dec 2004 20:47 GMT
Louis,

If I understand your question properly, you're interested in how to
represent a hierarchy in a dimension table, I assume without snowflaking
your schema?  How deep is the hierarchy?  One possible solution may be to
create additional columns in the dimension table such that each row will
contain a cross-tabulated representation of the hierarchy.  That will
generally solve the problem easily -- and perform well -- if the hierarchy
is shallow enough (and its depth is known in advance).  If you can't make
those kinds of guarantees during the design phase, I'm not sure how to solve
the problem other than snowflaking and having a separate hierarchy table.

Signature

Adam Machanic
SQL Server MVP
http://www.sqljunkies.com/weblog/amachanic
--

> For example, say I have a purchase fact table.  Then I have a customer
> dimension.  In the customer dimension, we have city, state, name, etc, and
[quoted text clipped - 15 lines]
> Also, if there is a good way to search for this kind of discussion on
> groups.google.com, advice there would be graciously accepted :)
Louis Davidson - 14 Dec 2004 04:21 GMT
Actually I was asking without any preconcieved notions.  I am very new to
the whole process of building a data warehouse so I am just learning.   I
subsequently found the very same answer in Kimball's book, which I
apparently missed the first time I read through it.  Being that I have built
only OLTP systems for 10 years, it is still a really strange process to me
:)

Signature

----------------------------------------------------------------------------
Louis Davidson - drsql@hotmail.com
SQL Server MVP

Compass Technology Management - www.compass.net
Pro SQL Server 2000 Database Design -
http://www.apress.com/book/bookDisplay.html?bID=266
Note: Please reply to the newsgroups only unless you are interested in
consulting services.  All other replies may be ignored :)

> Louis,
>
[quoted text clipped - 40 lines]
>> Note: Please reply to the newsgroups only unless you are interested in
>> consulting services.  All other replies may be ignored :)
Adam Machanic - 14 Dec 2004 05:19 GMT
> Actually I was asking without any preconcieved notions.  I am very new to
> the whole process of building a data warehouse so I am just learning.   I
> subsequently found the very same answer in Kimball's book, which I
> apparently missed the first time I read through it.  Being that I have built
> only OLTP systems for 10 years, it is still a really strange process to me

Wow, that's a great boost to my ego (giving the same answer as Kimball) -- I
haven't read his book in several years so either I absorbed it really well
or I'm finally really getting the hang of data warehousing :D

I am actually working on my first "real" OLTP system at the moment, after
years of doing DW (and a few very small semi-OLTP systems, nothing like the
retail system I'm working on now), so I'm in the opposite boat... Finally
learning how to properly deal with deadlocks, etc.  Fun stuff!

Signature

Adam Machanic
SQL Server MVP
http://www.sqljunkies.com/weblog/amachanic
--

 
Sign In
Join
My Latest Posts
My Monitored Threads
My Blog
My Photo Gallery
My Profile
My Homepage

Start New Thread
Enable EMail Alerts
Rate this Thread



©2008 Advenet LLC   Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.