Apply rules and validate data

Use business rules to measure the quality, completeness and accuracy of your data.

A rule is a Function that returns a true or false value, which translates to a pass or fail. This can be either a default Function (like 'Contains' or 'Is Date') or a custom one.

Suggest validation rules

The easiest way to get started is to use Suggest validation rules, which profiles the data to then suggest formulas to be applied to columns as validation rules. Uncheck any suggestions that are not useful and apply to create rules separated into rule groups ready for further changes if required.

In Explore mode, Profile your data to see the Suggest rules action button, which will create a new Workflow containing a Validate step.

Creating rules and groups

On the Validate step, click Rules to view and edit any existing rules and to Add rule or Add group.

Each rule has a name and summary, a parent rule group and a Function definition. The Function can either be a pre-existing Function in Data Studio, or a custom Function created in the Function builder. When building a custom Function, pick one of the relevant columns (or a parameter) as the input value and ensure the Function returns true/false as the result.

Rule groups represent logical collections of validation rules and can have different thresholds set. For example, a group of compliance rules may have a fail threshold of 99%, but a data quality group of rules may use a lower threshold and be increased over time as data and processes are improved.

The Status result column for each rule is based on these thresholds, so Red below the fail threshold, Green at or above the pass threshold, and Amber in between.

Each group has a name and description, pass and fail threshold, plus an optional column to weight the results. As well as counting the number of Passed rows and Failed rows, the results will also include Passed weight and Failed weight columns, which contain a sum of the chosen column for all the passing/failing rows. For example, weighting each rule's results by total sales allows you to prioritize data quality issues for the highest spending customers with the most overall impact.

Viewing results

Validation results are available in several formats:

• Show passing rows contains the rows that pass all of the defined rules.
• Show failing rows contains the rows that fail at least one validation rule.
• Show all rows is a combination of the two above, comprising the entire input data alongside the rules and a Pass/Fail result for each.
• Show results by rule summarizes/aggregates the results for each rule with a count of passes and fails.
  • Show results by rule for analysis is the same data, un-pivoted into rows.
• Show results by group is similar to the above, but useful for a higher level view of aggregated results when there are many rules.
  • Show results by group for analysis is the same data, un-pivoted into rows.

Similar to the Source and Profile step, the Validate step has a Source metadata dropdown that enables you to include the lineage metadata of input data. This information (such as file name or batch ID) is useful to include as additional columns in the aggregated results.

This step validates and enriches addresses in bulk.

Addresses will be cleaned by verifying them against the official postally-correct address files for the relevant country. Cleaned addresses are assigned a match result, based on the accuracy of the original address. You can define layouts specifying the number, content and format of the address output columns. Choose one of the available Additional datasets to enrich your data. The datasets that are available to you depend on your license.

Validate emails based on the format or domain address.

Select the Email column and pick one of the two available Validation type options:

• Format Check: Checks whether the value matched a valid email format. Returns either true or false.
  Examples of valid and invalid email formats:

  Format	Result
  info@gmail.com	Valid
  first.second-name@gmail.com	Valid
  first.name+tag@gmail.com	Valid
  name@info@gmail.com	Invalid
  name"not"right@test.com	Invalid
  another.test.com	Invalid
  name@incorrect_domain.com	Invalid
  com.domain@name	Invalid
  first_address@gmail.com, second_address@gmail.com	Invalid

  Only one email can be validated at once; lists of emails as seen in the last example will be rejected.

• Domain Level: Checks whether the value has a domain that exists and is an email server. This option returns both an overall validity result (true or false) in the Email domain: Result column, and additional information in the Email domain: Error column describing the reason for failure. The possible outcomes are:

  Error	Result	Description
  Valid	True	Domain exists and is a mail server.
  Bad format	False	Email format is invalid.
  Invalid domain	False	Domain validation check failed. The domain may not exist, or may have been flagged as illegitimate, disposable, harmful, nondeterministic or unverifiable.
  Invalid name	False	Local part validation failed. For example it may have been identified as a spam trap or role account such as "admin@server.com".
  Failed to perform DNS lookup	False	An error occurred when attempting to perform the DNS lookup.

Domain level validation results are cached with results refreshed every 30 days. The cache validity is configurable in Settings > Workflow steps by changing the Email validation cache validity setting.

Click Show step results to view the results.

Validate global phone numbers using the Validate phone numbers step in Workflows.

Connect the step to the source step and specify the following:

• Select phone column - specify the column containing the phone numbers you want to validate
• Select country - pick the country to use for phone validation. You can either select a single country or pick a column with country data. For the latter, please ensure that the country names adhere to ISO 3166-1 alpha-2, ISO 3166-1 alpha-3 or ISO3166 country name standards.
• Select validation type - choose the validation rule(s) that will be applied:
  • Valid phone: shows True for phone numbers that have been successfully validated against the selected country and False otherwise.
  • Valid phone region: shows True for phone numbers that have been successfully validated against the region of the selected country and False otherwise.
  • Possible phone : shows True if the phone number could be a valid phone number for the selected country based on its format but with a lower confidence than the Valid phone selection. It shows False otherwise.
  • Invalid phone: shows True for phone numbers that have been unsuccessfully validated against the selected country and False otherwise.
  • Invalid phone region: shows True for phone numbers that have been unsuccessfully validated against the region of the selected country and False otherwise.
  • Not possible phone : shows True if the phone number could not be a valid phone number for the selected country based on its format but with a lower confidence than the Valid phone selection. It shows False otherwise.

Click Show step results to view the results. The following columns will be appended to your data:

• Validation results – shows the result of the validation rule (one column per each applied rule).
• Phone Country – shows the country for the validated phone number.
• Phone Number Type – shows the phone type (e.g. mobile or fixed line).