Building Making It Happen
Building Making It Happen
  Sign-in         Register
    
Principles and Rules Listing Page
Handling Sparse Dimensional tables
In case you have a dimensional tables where significant proportion of instances have majority of the fields as null data in the records, one should look at creating a snowflake to save diskspace, incase the database system is a fixed field records (where null fields also occupy disk space).
 
This page of 'Principles and Rules' is linked to:  Data Warehousing, Data Analysis/OLAP, BI platform Tools Evaluation, BI business intelligence end-to-end view,

Let's assume that as an organization, you are having wide variety of products, ranging from very simple products (needing say 5 attributes) to complex products (needing 50 attributes). If you are selling mostly simple products (say 90% by volume), the product dimension table will be a sparse table, with most of the cells as Null. As long as you have a variable field database (like Oracle), it's OK. However for a fixed field database, you will need to have a way to reduce the storage space over-head. The best way to handle it is to have snow-flake whereby the common fields (which are populated for simple as well as the complex products) are in the main dimension table and the ones which are specific to the complex products are in the snow-flake table linked to the main dimensional table.

   Access more details on this page   
Quick Feedback- Was this information helpful ?
Relevant Links to this page
TOPIC - Data Mart Business Theme Matrix in Data Warehouse Dimensional Model → Principles & Rules → Dimensional model has to be aligned to the Entity-Relationship → Principles & Rules → Always Use Conformed Dimensions → Principles & Rules → You may not be a able to have a perfect ETL → Practice Techniques → Handling Sparse Dimensional tables → Principles & Rules → Do not separate the parent and child line item data → Practice Techniques → Managing time-stamps across multiple time-zones → Practice Techniques → Recording events in multiple currencies → Practice Techniques → Handle different units of measure in the same fact table → Principles & Rules → Handling of Null foreign Keys in fact tables → Principles & Rules → Dimension Attributes as NULL → Principles & Rules → Don't rely too much on Meta Data Tools to enforce Business Intelligence → Principles & Rules → Don't wait for universal models for Data Marting → Principles & Rules → Add extra buffer for ETL phase → Principles & Rules → Homework before interviews is must (Business Requirements Phase in Data Warehouse) → Principles & Rules → Excel is the competition, which should be challenged → Principles & Rules → Avoid Pure MOLAP → Practice Techniques → Field Tips Series- Streamlining & Cost-Reduction in Business Intelligence- Consolidate Data-Marts → Practice Techniques → Field Tips Series- Streamlining & Cost-Reduction in Business Intelligence- Licensing & Maintenance Contracts → Practice Techniques → Field Tips Series- Streamlining & Cost-Reduction in Business Intelligence- Governance & Standards → Practice Techniques → Field Tips Series- Streamlining & reducing cost of Business Intelligence- Evaluate Open Source → Principles & Rules → Master Data Management- Making a Right Start → Practice Techniques → How to integrate stand-alone BI environments- Gradual Approach → Principles & Rules → Business owned applications are a reality- Manage it → Principles & Rules → New Data Standards- What about existing data and applications? → Principles & Rules → Handle Each Time-stamp in the Fact Table as a separate dimension → Principles & Rules → Keep Aggregates and Details data in different Fact tables → Principles & Rules → Some considerations for Infrastructure in Data Warehouse → Principles & Rules → For Core BI platform go for a single, established and robust player → Principles & Rules → Don't be guided only by the business requirements for your Business Intelligence → Practice Techniques → Using Synonyms and Views → 
 
Back
Featured Pages
Excel vs. BI
Data Quality Assurance Track
Don't rely too much on Meta Data Tools
Parallel Dimensional Hierarchy

Make 'Executable' Strategy
Maximize Results
Maximize People
Manage Execution

Featured Pages
Dimensional non Strict Hierarchy
Enterprise Intelligence' Evaluation- Data Pipelines
Vendors in Data Quality Program
New Data Standards on existing apps
About Us |  Contact Us |  Privacy Policy |  Terms and Conditions |  Copyright |  Disclaimer |  Site Map
  Executionmih.com © 2010 All rights reserved | Owned by Sagacious Management Solutions | Powered by : Giga Soft Systems Pvt. Ltd