Conformity metrics

Conformity metrics let you specify one or more conditions and check the percentage of rows that conform to the condition.

Create a conformity metric

  1. In the Explorer tree, select the data asset that you want to measure.

  2. On the asset's Explorer menu select + Create Metric:

  1. In the metric configuration form, for Metric Type select Conformity.
  1. Enter a Metric name.
  2. By default, new metrics go Live shortly after you create them. If you want the metric to start off paused, under Metric status, set Status to Paused.
  3. If desired, change the value of Quality dimension. Each data quality metric measures one dimension of data quality: accuracy, completeness, timeliness, or (if none of these applies), custom. You can sort and filter your metrics by dimension, and you can view them summarized by dimension in a data quality dashboard.
  1. Select Next. Step 2 (Configure Metric) appears in the main form:
  1. Under Select data asset, make sure the right data asset is selected.
  2. To the right of Conditions select Add + to build the condition that checks values for conformity.
  • The selections available vary based whether you choose a single column of type string, a single column of type numeric, or multiple columns.
  • When you create a single-column conformity check, it appears at the column level in your Explorer tree. Currently one multi-column metric is available, "Is Unique", and that will appear at the table level in your Explorer tree.
  • Depending on the datasource, if you use single-quoted string constants you might need to escape the backslash character in any backslash-sequence. See the datasource vendor's documentation for details.
  • If you add more than one condition, the second and any subsequent conditions must all be true to conform.
  • To delete a condition, use the waste bin icon beside it to the right.
  1. Perform any optional configuration (see below), then select Next to proceed to Step 3, where you can preview and save your metric.

Optional configuration

  1. Add a WHERE clause.
  2. If needed, override the inherited table configuration.
  3. Add a slice column.
  4. Add seasonality.
  5. Adjust the backfill duration.
  6. Choose a missing value filling.
  7. Add a failing records query.
  8. Add tags.

Step 3 (Preview)

Once you have configured your metric you can preview it to view a live query without creating a datapoint. If the preview doesn't work the way you want, you can select the tab for a previous step and modify the configuration as needed.

  1. On the Step 3 (Preview) tab, set the date range for the preview, then select Preview.
  2. When you're happy with the result, select Next to continue.

You don't have to preview your metric— you can just click Next to proceed to Step 4.

Step 4 (Related metrics)

You can add related metrics to help streamline your analysis of any incidents identified by the metric. For steps, see Related metrics.