Browsing All posts tagged under »Dimensional Modeling«

Process Modeling – Is this a fact or a dimension?

November 23, 2011

8

This is a question I see a lot on forums (and also get asked) and it’s generally based on confusion around events and processes (such as an incident, a bank application process etc). Let’s remember what the difference is between a fact and a dimension. A fact table stores a business process.  A fact is therefore an […]

Tracking History with Slowly Changing Dimensions

September 24, 2011

0

In this post I’d like to show a few of the different ways to maintain history. As anyone who has been in data-warehousing for a while can attest to, the two most common scenarios that business users want to see are the data as it was (Kimball Slowly Changing Dimension Type 2) , or as-is […]

Dimensional Modeling Worst Practices

May 16, 2011

2

As a consultant I see a lot of BI/DW solutions with various clients, and am often called in to fix performance issues. Often times developers/architects will ignore the Ralph Kimball methodology, or try to “improve it” by breaking a few rules. So I thought it might be worthwhile for me to post some of the […]

Data Modeling Conundrum Follow up

March 17, 2011

9

As a follow up to my previous post on a Data Modeling Conundrum in January (see https://jsimonbi.wordpress.com/2011/01/16/data-modeling-conundrum/) I left it open as to what the appropriate dimensional model should be. To recap from my first post, let’s say that I’m trying to track Marine Incidents. Zero or many vessels may be involved in an incident, […]

Populating a many-to-many bridge table pt2

February 1, 2011

1

I have been testing the previous example code (see https://jsimonbi.wordpress.com/2011/01/25/populating-a-many-to-many-bridge-table/) to populate a many to many bridge table with a large dataset and wasn’t happy with the performance. So I’ve  gone back to the drawing board on how to do this. I wanted a set based solution, and something that was both simple and efficient. To recap, if […]

Populating a many-to-many bridge table

January 25, 2011

2

STOP! For a simpler, faster and more elegant solution to the following problem, see this: https://jsimonbi.wordpress.com/2011/02/01/populating-a-many-to-many-bridge-table-pt2/ In a recent post on dealing with the many-to-many situation (https://jsimonbi.wordpress.com/2011/01/16/data-modeling-conundrum/) , I illustrated the use of a bridge table in a dimensional model, as prescribed by Ralph Kimball. So if for example we are recording sales, and many […]

Data Modeling Conundrum

January 16, 2011

4

A post on the Kimball forum found at www.rkimball.com has raised a question very similar to one I worked on five years ago when I implemented an Oracle OLTP system capturing Marine Incidents, and the associated data mart for reporting. The question was this: How to model an incident when an incident may involve zero or many […]