Understanding Data Quality Evaluation Results – Or: More than just OK and Not OK
With SAP Master Data Governance you can evaluate the data quality of your master data using custom rules. Even though the outcome of a single data quality rule is either OK or Not OK, the overall picture of a complete data quality evaluation is a little more diverse.
Each data quality rule has a property called base table. This is the database table used as the basis for the evaluation of the rule. Examples:
- If the base table of a rule for product master data is Basic Data (MARA), the system will evaluate this rule for each product once.
- If the base table is Plant Data (MARC) the system will evaluate the rule for each product depending on the number of plants for which the product was created. This could be once, multiple times or even none.
The scope of the data quality rule defines to which master data the rule shall be applied. For all data in the scope of the rule, the rule’s condition is evaluated, and the evaluation of the rule produces the Outcome of either OK or Not OK. For all data that is not in the scope of the rule, the outcome is undefined.
Scope expression of the rule above:
Condition expression of the rule above:
The system stores the evaluation outcome per row of the base table. This entire information of outcomes of rules and base tables is called Evaluation Results. You can view the evaluation results per rule or per base table. For example, the evaluation results of a rule are available in the header of the data quality rule. The section Evaluation also tells you how many rows were available in the base table, how many of these passed the scope expression and were actually checked.
There are 3 cases when there are no evaluation results of a rule available:
- The rule is not active and/or it was not yet considered in any evaluation
- The rule was evaluated, but according to the rule’s scope there was no data that produced any outcome
- The rule was evaluated and evaluation results were available, but these results were already deleted (for example from the app Manage Data Quality Evaluation)
The app Manage Data Quality Evaluation for Products shows you for each evaluation that was completed how many product master data records were evaluated (here: 5.627 products) and how many Evaluation Results are available (here: more than 145 thousand).
Like for a single rule the available evaluation results of an evaluation can be 0. Typical reasons are:
- The results were deleted using the Delete Results action
- The system automatically deleted the results due to the housekeeping settings of a subsequent evaluation
Looking at the details of the evaluation, you see how many of the 145k rule evaluations had the outcome OK (62.8k) and how many had the outcome Not OK (82.8k). You also find a chart that shows you the results per outcome and base table.
You can use one of the Evaluation Results apps to analyze the evaluation results. For example, the app Evaluation Results for Plant Data of Products provides you with the view on plant-level: it displays results of rules with the base tables Plant Data (MARC) and Storage Location (MARD) in one single table, one row for each combination of product, plant, rule, and outcome. Because there can be results for more than one storage location for such a combination, there is one row for the results with outcome OK and one row with the results with outcome Not OK. Such a row is called evaluation item. As an example, the screenshot shows the results for product ZAS_BAT1 in the table. The last two rows show the results for the rule STPIT1MARD. There are two storage locations that fail the condition of the rule and there is one storage location that passes the condition of the rule.
As you see, there is more than just the binary outcome of a data quality rule.