Even though data warehouses are best implemented using an interative project management methodology, all of the tried and true testing disciplines still apply.
Unit, string/integration, user acceptance and production/volume stress testing phases apply for the ETL modules and BI modules you design for your solution.
The focus of the testing is more data intense than operational intense. The goal of the ETL testing is to provide you accounted for every data element you processed from the source and placed in the target. Simple before and after SQL is an effective way to prove row level processing and column level processing for numeric oriented columns (sums, averages) and "code columns" (distinct counts by code values). You still need to explicitly test your ETL logic and BI reporting/dashboard logic, so make sure you have a spec and develop test plans that try to prove you both didn't break the spec (negative) and that you indeed met the spec (positive).
Dig Deeper on Oracle data warehousing
Related Q&A from Mike Lampa
When trying to design a data warehouse, we often try to model the database on the operational data model. Are there any guidelines in trying to ... Continue Reading
What is a surrogate key in a table? Continue Reading
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.