Users can optionally organize test rules into data quality dimensions to surface specific data quality issues in the dataset. Each data quality dimension generates a data quality score, enabling users to immediately learn what types of data quality issues exist, how prevalent they are, and which fields have each issue.

There are six dimensions to choose from, each representing an essential characteristic of quality data: Accuracy, Completeness, Consistency, Timeliness, Uniqueness and Validity.

To configure data quality dimensions, users associate a test rule with a dimension when they configure the rule (by setting the Dimensions property). Users can associate test rules with the dimensions of their choice. Test rules are Assert, CompareToConstant, FuzzyMatch, IsNotBlank, IsNotDuplicate, IsNotNull, InRange, MatchesRegex, RemoveDuplicates and RemoveDuplicatesFuzzyMatching.

For an example of how dimensions work and the visibility users gain from using them, see Example: Data Quality Dimensions.

Once data quality dimensions are configured, users manage them on the Configuration tab in the Dimensions section (see Managing Data Quality Dimensions), and view output results in the Statistics tab (see Viewing Statistics).

The following are the dimensions and the data quality characteristic they represent:

Dimensions also enable users to specify the importance levels of rules by assigning rule weights, which are figured into dimension scores. Rule weights are useful, for example, to call out business critical fields.

A dimension score is derived by aggregating individual rule pass results (fields that passed test rule criteria) using the following calculation:

(Rule PASS Percentage * Weight)/Total Weights

For details about how scores are calculated, see How Data Quality Dimension Scores are Calculated.

All scores are updated upon profile execution and visible in the Statistics tab (which opens automatically after a profile is executed). See Viewing Statistics. If no dimensions are configured, dimension scores are not generated.

To disable dimensions, move all rules out of dimensions.

Refer to the following topics for more information: