Jackie McDougall - Learning to Love Data - EuroSTAR 2010
-
Upload
eurostar-software-testing-conference -
Category
Technology
-
view
43 -
download
0
description
Transcript of Jackie McDougall - Learning to Love Data - EuroSTAR 2010
Unissons nos Talents
T O G E T H E RT A L E N T E D
1
Jackie McDougall
07894 607776
Learning to L♥ve Data Warehouse Testing
2Learning to Love Data Warehouse Testing
Investigating Data Warehouses
Project Background
Testing Approach
Client plan
IdealMy Initial Thoughts
My View
Lessons Learned
The Way Forward
Agenda
3Learning to Love Data Warehouse Testing
Investigating Data Warehouses
THE TRUTH IS OUT THERE …
4Learning to Love Data Warehouse Testing
What is a data warehouse?
“A warehouse is a subject-orientated, integrated, time-variant and
non-volatile collection of data in support of management’s decision making process” Source: “What is a Data Warehouse?” W.H. Inmon, 1995
“A data warehouse is a copy of transaction data specifically structured for query and analysis”. Source: “The Data Warehouse Toolkit” R. Kimball, 1996
A Data Warehouse is NOT a traditional relational database…
What’s the difference?
Investigating Data Warehouses
5Learning to Love Data Warehouse Testing
UK Public sector client
Government directive to achieve certain
heat targets by next year
Solution was going to be ETL of data
from existing sources to single Data
Warehouse
Third party providing warehouse
solution
Reports to ascertain if/when threshold
levels reached/breached
Project Background
6Learning to Love Data Warehouse Testing
Testing Approach – the Client’s Plan
Ref Data &
Data Load Test
18th Dec – 6th Jan
Dec
2009
January
2010
March
2010
April
2010
February
2010
Warehouse
Implemented
14th – 17th Dec
Pathway
Matching
14th – 20th
Validation Test
7th – 13th
Test Preparation
9th Jan – 27th Feb
Redevelopment
28th Jan – 3rd Feb
Archive DL & Report Testing
3rd Feb – 9th Mar
Real-time DL & Report Testing
17th – 30th Mar
Redevelopment
10th – 16th Mar
User Report Testing
31st Mar – 21st Apr
Key: Dev Team
IT Project Team
Info Dept
Users
Live
26th Apr
7Learning to Love Data Warehouse Testing
Testing Approach – Ideal
ETL Target
(DW)Reports
Source
Data
Staging
Data
•Count records
In source files•Count records in each
part of staging
•Count records
In target (DW)
•Count target
displayed in reports
E T L U
•Check all fields (directly extracted and
derived/calculated)
•Check all fields
displayed in reports
3rd Party Supplier
System Test
8Learning to Love Data Warehouse Testing
Test resources
Test Manager brought in
Engaged with other users
(Data Quality staff, Info
Dept testers, (beta)
testers)
Formal test strategy
What Environment
What Data
Inclusions/Exclusions
Resources
Acceptance Criteria
Stripped back test scripts
Release Process
Defect Management
Risks and Assumptions
Testing Approach – My Initial Thoughts
SIT – ETL Requirements Testing Risk Based Testing Scheduled regression
testing
SIT – reports Exploratory Testing
UAT – ETL Reconciliation Acceptance criteria
UAT - reports early previews
9Learning to Love Data Warehouse Testing
Testing Approach – My View
ETL Target
(DW)Reports
Source
Data
Staging
Data
E T L U
Perfect world testing…
•Sample check key
fields for calculated/
derived data
•Sample check data load
of transaction data and
reference data
•Sample check
data validation/
error process
•Reconciliation
Project Team
Pre-UAT/
Sanity check
•Reports
User Acceptance
Testing
r
TOP 10 COMMON PATHWAYS
(RISK BASED/
SCENARIO APPROACH)
10Learning to Love Data Warehouse Testing
Testing Approach – My View (SIT)
11Learning to Love Data Warehouse Testing
Lessons Learned
Define your quality entry criteria that you expect from your supplier
Insist your development team/supplier provides adequate
design/development documentation
Get the requirements right (or at least fixed!) and test them
Consider data quality and when it gets addressed
Get a test strategy (a real one)
Get the right (skilled) resources on board (early)
Think carefully about the reconciliation exercise
Consider how (the project team) will support the solution going
forward
12Learning to Love Data Warehouse Testing
The Way Forward?
Create a DW testing ‘companion’ – crib sheet for fellow users of things to think about at different stages
System Test/SIT – examine how I can formalise this with colleagues and use techniques like requirements testing, RBT, exploratory testing to help them…
UAT – promote the Top 10 common pathways approach (combination of RBT/scenario testing)
Learn more about DW and BI
Phase 1 Testing? – still ongoing
Independent Review of solution to aid Phase 2 planning
Established test & supporting processes will be maintained
Re-use of test scripts and data possible
Potential for use of test tools in the future
13Learning to Love Data Warehouse Testing
Thank You
Advice for me or questions?
Jackie McDougallSenior Test Consultant
Testing Services Group
Carnbroe House, Finch Way
Strathclyde Business Park
Bellshill, North Lanarkshire
ML4 3PR
Tel : +44(0)1698 574249
Mob:+44(0)7894 607776
Mail : [email protected]