Functional Dependency Discovery
Functional dependency discovery provides information about dependencies between pairs of columns in a data source.
A pair of columns are functionally dependent if the values in one column can reliably predict the values in another column. For example, if a dataset contains an Employer ID column and a date of birth column, the date of birth should be the same in all rows that contain a given Employer ID.
Functional dependencies can highlight potential data quality issues by identifying records the records that do not conform to a column functional dependency. For example, if 99.8% of rows in a data source are functionally dependent, there is a high likelihood that the remaining rows may contain incorrect information.
Functional Dependency Inference Properties
The Functional Dependency Profiling view provides information about the functional dependencies between columns.
The following table describes the functional dependency inference properties in the Functional Dependency Profiling view:
Property | Description |
|---|
Override the default inference options | Allows you to configure custom settings for functional dependency inference. |
Max Columns in Determinant | Number of columns that the profile can combine to find a determinant. |
Max Rows | Number of rows to profile. |
Dependencies Returned | Number of dependencies that the profile displays. Default is Minimum Coverage, which displays the smallest set of dependencies where each column appears at least one time in a dependent or determinant. |
Max Dependencies Returned | Maximum number of dependencies that the profile displays. |
Conformance Criteria | The minimum percentage or the maximum number of rows for dependency violations that the profile allows when determining functional dependents. |
Inferred Functional Dependency Properties
After you run a single data object profile, you can use the Functional Dependency Inference view to view the details of inferred functional dependencies between columns in the data source.
The following table describes the inferred functional dependency properties in the Functional Dependency Inference view:
Property | Description |
|---|
Determinant Columns | Name of the column analyzed for functional dependencies. |
Dependent Columns | Name of the column dependent on the determinant column. |
% Null | Percentage of null values for the column. |
% Conforming | Percentage of functional dependency match. |
Verified | Determines whether or not the columns are functionally dependent. |
Last Run Time | The date and time that the functional dependency profile last ran. |
Functional Dependency Violations Properties
The view provides information about the functional dependencies between columns. After you run a single data object profile, you can use the Functional Dependency Inference view to view the details of the functional dependency violations in the data source.
The following table describes the functional dependency violations properties in the Functional Dependency Inference view:
Property | Description |
|---|
Determinant column | Name of the column analyzed for functional dependencies. |
Distinct Dependents | Number of unique functional dependencies. |