You can edit an existing semantic type in Talend Dictionary Service to impact how your data is validated in Talend Data Preparation.
Predefined semantic types in Talend Data Preparation are based on standard values, but you may need to tailor them to match your own data. Some data that you would expect to fall under a predefined category, may be considered invalid.
Let's take the example of a dataset containing a list of customers, with their email addresses, date of birth, and the country they live in. You can notice that all the entries for America are considered invalid. While it is indeed not a valid country name, it is the value that your company is using and you would like to make it a valid value.
The problem here is that America
is not one of the expected value for the
semantic type in Talend Data Preparation. The valid
entry in this case would be United States or
United States of America.
To avoid having this problem in the future, you will update the
country semantic type in Talend Dictionary Service, and add America to the list of valid entries. The change will be
automatically available in Talend Data Preparation.
- Open the Semantic types view from the left panel of the Talend Data Preparation homepage.
From the list of existing semantic types, click the
Country type to open it.
In this window, all the parameters of the semantic type can be modified, including the list of entries used to discover or validate data.
- In the Values list, point your mouse over the United States entry and click the pen icon that is displayed on the right.
- Right after United States, enter America as a new value, separated by a comma.
Click the check icon to validate your change.
All the comma-separated values that are in the same row, are set as synonyms. As a consequence, America will now also be considered a valid value for the
Click Save and publish to propagate the change in
Talend Dictionary Service and
make it available to the Talend Data Preparation users.
The change in semantic types is instantly effective in Talend Data Preparation for every new dataset that you import. For existing datasets, you need to duplicate the column or reimport your dataset.
- Go back to your dataset with the column containing the customers countries.
Duplicate the column with the updated semantic type applied,
Country in this case.
You can see in the quality bar under the column header that there is no invalid values anymore.
country semantic type has been manually updated to support a new
From now on, when dealing with data that are matched with the
country semantic type, America will be considered a valid value.