SAP BW – Automated Test Tool

[Martin_Kuma]

Even simple changes in reporting or planning interfaces could lead to unexpected dependency issues. Testing by the end-user should be the last Option. Below can be found an Automated Test Tool (ATT) using the approach: Test-Fail-Build-Test.

In BW the front-line object is the Query (mostly). APD, OpenHub, … are not considered as the expectation is that BW is used for reporting or planning, not as a EAI system.

This automated test uses the Query run-time interface to fire the Queries in the background and save results into an ADSO.

ADSO is used to process the results with BW tools.

The Tool can also be used for:

• pre-caching of queries. For example: performance boost for the first user after cache invalidation. Queries can be fired with pre-defined selections, so the tool can also be used for pre-caching.
• monitoring of KPIs. For example: unexpected KPI changes caused by changed processes in the source systems

Objects Used:

Object	Description
Query Executor	Using RSRT functionality to fire the queries in the background
BW Reporting objects	BW objects to save and evaluate the results
Results Processor	Class with implemented methods to process the results

BW Versions: 7.4, 7.5, BW/4 (Expected EndOfLife: 2040)

Example how to use the tool: / Tool was used in the following case:

Let’s say we have a Corporate Controlling map of 100+ KPIs with deep dependencies. Requirement is to adjust a couple of the KPIs based on the latest changes in the KPI repository (for example ROCI or NWC). Such Query uses a number of cell-based formulas. We also have dependencies across multiple queries.

Approach would be:

1. Get the queries currently used by the Corporate Controlling department (usually the number of Queries is rather low, but Queries are complex)
2. Get a general selection for the queries (ideally a high-level selection to bypass any local accounting specialties). Query-Selections are not in the new Statistics, thus have to be asked for. Alternative is to enhance the Statistics to have the Query-Selections saved as well (not in this article however)
3. Execute the Queries with the Selections via the Automated Test Tool (Version/Variant-BeforeChange)
4. Do Implement the changes
5. Execute the Queries with the Selections via the Automated Test Tool (Version/Variant-AfterChange)
6. Compare the KF of the two Versions/Variants. All differences must be explained / traced

How-To:

First, we need to build the Query Executor. We use the RSRT interface functionality. RSRT interface is a stable interface used in BW/4 as well.

The Query Executor will save all results into ADSO which is consumed via HCPR.

Class specific methods for evaluation. The main evaluation method would be: compare.

Code WIP