Fact table structure
I am pretty knew to the matter and need some guidance to the below. I have been of the view that facts shouldn't contain any descriptive information. But, pretty recently i switched companies and found that we have a scenario where facts in the area had

character data in a fact table?
I just changed companies. i was accustomed to pure star schemas with fact tables containing only foreign keys to dimension tables and number measures that could be aggregated (sum, average, min, max). At my new job, I find dims with foreign keys

Textual Facts in the fact table
Its a new financial 'capital markets' dimensional reporting model (for IPO - raise money for an issuer by issuing new equity common stock securities) that's in design stage. There are some textual attributes of the grain of several fact tables for which

Orders Fact Table
I've been reading more and more Kimball literature and I'd like clarification on something. Transaction Fact Tables are rarely revisited because they do not need to be updated, deleted, etc but typically only appended to. He goes on to describe

One Fact table having records at different granularity level
Hello, My question is just validation for an approach that we followed for one of our client requirements. Based on the requirement the FACT table should be constrcuted by the records from 2 Source table A and B. The structure of the table is as

fact table as a dimension
Hi I'm presently modeling a data mart and I would like to use a fact table as a dimension for another fact table. Is it something doable and efficient or is there a better way to do that? if yes should i just create a surrogate key and link both

Fact Table Indexing Strategy
Please explain the indexing strategy which is in use in your organisation for fact tables in your current database(oracle, sybase, db2) Specify volume as the indexing strategy may be based upon the volume of data. Is there a well laid out strategy

Fact table with non-numeric measure
I have a transactional fact table containing results of inspections. The results are non-numeric, but do represent a measurement, that is inspection result (like "pass", "fail", "pass with infraction", "failure with

Loading Fact table
Hi, I have following tables FACT, A_DIM, B_DIM. How do I update relationship between a_key and b_key in the FACT table when new record arrives in TYPE2 a_dim dimension? Below is simple create scrip to illustrate what I mean. create table fact (a_key

dimension model for many indicators field
I have 30+ fileds that has either 1 or 0 ( true or false)value in a fact table Patid -- -- --- PTInd_1 Ptind_2 Ptind_3........ PTind_30 -- etc 1

Fact Table Grain at a Sub-Atomic Level
I'm thumbing through the Design Tips and would love to find some thought given to the statement made in The Data Warehouse Toolkit: The Complete Guide to Dimensional Modeling (pg 34): "Preferably you should develop dimensional models for the most

Cost plans for projects - One fact table or several fact tables?
Hi, I'm currently struggling with my Dimensional Model. It is about projects with their related cost plans. Every project starts with an initial cost plan (called WIPL), the project is saved as version 0 and the WIPL attribute and the ACTUAL attribute i

DatawareHouse modeling (factless table)
Hi everyone ! (To begin, I want to apologize for the bad english i'm gone use cause im french canadian) system i'm using : Sql Server 2005 Ent SSAS 2005 (For cube) SSIS 2005 (For ETL) I'm in the process of building a datawarehouse for the

Transactional detail fact table w/complimentary snapshot table. How do the two play together?
Take a detail daily banking transaction table with standard credit and debit level transactions. I want to build the accompanying snapshot that would present the higher level rollups of these lower transactions: There are about 250 daily transaction

NULLs in source lead to duplicates in Fact table
Hi, my historic source data has inconsistencies. I have a product with a 1:m to product_detail which has a 1:m to invoice which has a 1:m to invoice detail. Invoice_detail is the grain of my fact table, but some of the historic data has nulls in the lin