In Integration Services, a data cleaning methodology that examines values in a dataset and identifies groups of related data rows and the one data row that is the canonical representation of the group.