Datawarehouse testing

Does the data meet the specified business requirements and objectives?

Are there metrics for data quality? Who defines them?

Is there test (synthetic) data?

Does the volume of the test data correspond to the live volume?

Is the data complete or is some missing?

Does data tranformation actually enhance the value of data?

Are there forced error tests to check if and how the system can handle faulty data?

Compare OLAP results with data queried from the source database.

Organizations for which information (data) is a main asset for strategic decisions etc., need to assure its quality.

Are the schemas (ERD) understandable?

How easy is it to select, join, filter etc. the needed data?

Do access rights prevent the users from getting »their« data?

How long do various tasks such as ETL, joining fact tables to dimension tables etc take?

Is there a definition of a reference of

• data volume for ETL jobs and
• number of concurrent sessions and type of queries for normal workload?

Similar to performance tests, but it discovers bottlenecks under peak loads of data and heavy workloads of the system.

Can data be recovered after hardware failures or accidental misoperations (drop table...)?

How long does it take to recover the data?

How can it be verified that all and the most recently backed up data is recovered?

Who is in charge?

Is someone able to copy data without autorization?

What actions (DML and DDL etc) are audited?

Who checkes the audits? Interval of such checks?

Datawarehouse