LiveCompare 3.8: Smarter Impact Analysis

In the Introducing LiveCompare 3.8 post I said that we wanted to do a better job of analysing data changes such as IMG configuration changes.

First, let’s review what happens when we run Impact on a transport containing table content changes using LiveCompare 3.7R3.

LiveCompare examines all the table keys in the input transports to find the set of affected tables. LiveCompare then simulates a table definition change in order to feed the impact system a structural modification. LiveCompare then exhaustively discovers the links between the used objects and the (simulated) changing table definitions to find the impacted objects. Finally LiveCompare’s unique reduction algorithm identifies the most-at-risk objects to test.

The upshot of this is that if you were to change the configuration of a sales document type and run Impact, LiveCompare would most likely recommend VA01 (Create Sales Order) as the most-at-risk object to test.

But there’s quite a gap between VA01 and the configuration change we made. This is especially true if we have hundreds of test data scenarios for VA01.

How do we know which test data scenarios are relevant?

This is the question we’ve set out to answer in LiveCompare 3.8.

Table content changes are now processed using a new side-by-side comparison analysis that’s integrated in the Impact app. The result is a new Test Data worksheet in the report. The worksheet includes a list of all the affected tables with each table a hyperlink to a new side-by-side comparison report.

Each side-by-side table report looks like:

When LiveCompare detects that the affected table is part of the IMG it outputs a filtered view that includes only the customising activities that change the affected table.

Note: understandably, the IMG section of the report applies only to IMG tables. All other tables get the same side-by-side differences report, just no IMG section.

Now when we make a configuration change, LiveCompare will:

  • Identify the most-at-risk object to test.
  • Identify the test data scenarios that should be validated.

This combination delivers great results. Always 100% coverage for just 15% of the effort:

Things to note:

Impact must be configured with a Comparison System in order for LiveCompare to produce the side-by-side table key comparison report. Our use case is primarily to support custom releases where:

  • Analysis system -> DEV
  • Comparison system -> QAS

And run Impact before the transports are released to QAS. In this scenario:

  • DEV = TOBE
  • QAS = ASIS