Data rules (Watson Knowledge Catalog)

Data rules evaluate and validate specific conditions associated with your data sources by binding data rule definitions to physical data.

Data rules apply the logic specified in the selected data rule definition to the physical data. You bind selected assets to data rule definition variables. You can bind assets from more than one data asset by using joins. The output is written to a table, for which you can specify the content. After you create a data rule, you must run it to see the results.

For example, you want to make sure that each customer in your database has an address assigned. You select a data rule definition AddressExists, and bind column Address to the rule definition variable. After you run the data rule, you can see how many records meet the specified conditions, and how many don't.

You can create the output table with customized content for a data rule, or choose to display only statistics.

You can create a data rule either in the details page of a data asset, or from the data rule definition level. The data rule is saved in the data asset from which you selected data for bindings. If you use joins, the data rule is saved in each data asset from which you selected data for bindings.

Note: When a data rule contains joins, the quality score of each data asset is not impacted. To impact quality score of a data asset, the bindings in a data rule must come from one data asset.