You can create rules to validate the quality of the data in your sample. When created, these rules allow you to highlight exceptions to the rule to assist in building your data cleansing recipe steps.
NOTE: Data quality rules are not transformation steps. They assess the current state of the sampled data in the Transformer page. |
NOTE: As you apply transformation steps to the data, the state of your data quality rules is automatically updated to reflect the changes. If you delete columns or other elements referenced in the data quality rules, errors are generated in the Transformer page. |
You can add a rule from inside the Transformer page.
Steps:
If you have not created any rules, the panel is empty. To create a new rule, click Add rule.
Tip: You can review a set of suggested data quality rules that are based on your dataset and add them as needed. Click View suggestions. For more information, see Data Quality Rules Panel. |
Not Null
. See Examples below.May be missing: Some rule types support the May be missing checkbox. When it is enabled, the Data Quality rule allows missing values to be acceptable for a specified column.
NOTE: The May be missing rule parameter is not applicable to Not Null, Not Missing, Not Equal, Not In Set, and Formula rule types. |
Select the column or columns to which the rule applies.
Tip: Some rules can be applied to multiple columns. |
Click Add.
Review the previewed results.
Tip: To simplify the preview, click the Show Only Affected Columns checkbox in the status bar. |
The new rule is displayed in the Data Quality Rules panel. In the data quality bar for the rule, green indicates the row values that have passed the rule, and red indicates the row values that failed.
Tip: After creating a rule, you can jump back and forth between the Recipe panel and this panel to review how your changes to your recipe steps affect the data quality bars for your rules. |
Additional options are available in the context menu for the rule. For more information, see Data Quality Rules Panel.
The following data quality rule tests the values in the storeAddress
column to see if they are missing (empty) values.
The following rule evaluates the primaryKey
column to determine if all values in it are unique.
Suppose the values of your SKUs must be in the form of "SKU
+ 6 digits".
Following uses to perform the match. For more information on
, see Text Matching.
This rule tests the values in the orderColor
column to verify that all values are Blue
, Yellow
, or Green
.
In the following, the Acceptable values must be formatted as an array. See below.
You can add custom rules using formulas containing functions.
is the proprietary language used to transform your data. You can also apply the functions of the language to your data quality rules. For more information, see Wrangle Language.
Steps:
Formula
.In the Formula textbox, enter the formula to test your data.
NOTE: The formula that you provide must evaluate to |
For aggregation functions, you can group the evaluation of your rule based on the values in your grouping column.
Tip: You can group by multiple columns. The first column is the outermost grouping. |
To add the rule, click Add.
You can use data quality rules to perform some data analysis functions. For example, suppose you want to flag the dates where the total sales of all of your orders was less than 100.
When this rule is added, the rows whose date total is less than 100 are flagged in red.
To edit a rule, select Edit rule from the context menu for the rule in the panel.
To delete a rule, select Delete rule from the context menu for the rule in the panel.
When you generate a profile as part of your job results, you can download the profile in JSON or PDF format.
When you download the profile in JSON format, the set of rules for the job are also included. Search for profilerRules
in the JSON file.
For more information, see Job Details Page.
When flows are exported and imported, the rule definitions for the recipes in the flow are also exported. For more information, see Export Flow.