UNIQUESpalten

UNIQUESpalten is a data profiling methodology that identifies all minimal unique column combinations (UCCs) within a relational dataset. A UCC is a smallest set of columns for which the combination of values uniquely identifies each row. The technique is used in database design, data quality assessment, and privacy risk analysis to surface candidate keys and understand data dependencies.

The term combines the idea of uniqueness with the German word Spalten, meaning columns, reflecting its focus on discovering unique patterns across column sets. In practice, UNIQUESpalten serves as a practical preprocessing step for normalization, deduplication, and governance tasks by clarifying which column combinations can function as keys.

Methodology typically involves an iterative search over subsets of columns, with pruning based on known functional dependencies and row counts. Efficient implementations use hashing or bitmap indexing to test uniqueness quickly, and may employ sampling or approximation for very wide tables. Parallel processing is often applied to handle large datasets.

Applications and considerations include identifying candidate keys, detecting data quality issues such as duplicates, and informing indexing and normalization strategies. UNIQUESpalten also supports privacy risk assessments by revealing columns or combinations that uniquely identify individuals. Limitations include the potential combinatorial explosion as table width grows, sensitivity to missing values or inconsistencies, and the need for robust handling of nulls.

Example: in a customer table with columns such as CustomerID, Email, Name, Address, and Phone, UNIQUESpalten might identify Email as a UCC and, in some datasets, a combination like (Phone, Address) as another UCC. These findings guide key selection and data modeling decisions.