Mask sensitive information in unstructured files, such as PDFs and text documents, so that only authorized users can view it. Masking policies work on sensitive data that’s automatically tagged or identified.
Required Editions
| Available in: All Editions supported by Data 360. See Data 360 edition availability. |
How Unstructured Data Masking Works
Tagging is the
foundation for masking unstructured data. When Data 360 ingests unstructured files as
unified data lake objects (UDLOs), it analyzes the content and automatically applies
personally identifiable information (PII) tags based on recognized patterns. For example, an
email address such as mymail@example.com is automatically tagged with
pii.email. For the list of supported PII tags, see Einstein Trust Layer Region and Language Support.
After data is tagged, masking policies hide sensitive fields in query results. For example,
a support rep sees a customer’s purchase history, while details such as phone numbers and
email addresses remain masked.
You can tag UDLOs up to 5 GB per object. For larger data, split it into multiple UDLOs and create separate search indexes.
A single value can match multiple data types. For example, in mariarodriguez@example.com, the entire string matches an email address, while the substring rodriguez also matches the last name. In this instance, Data 360 fully masks the value to prevent accidental exposure. For example, mariarodriguez@example.com becomes XXXXXXXXXXXXXXXXXXXXX.
Unstructured tagging supports content in these languages:
- English
- Japanese
- German
- Spanish
- French
- Italian
- Configure Automatic Content Tagging (Beta)
Configure automatic content tagging for unstructured files and define a confidence threshold. Data 360 approves tags that meet or exceed the threshold, but others require manual review.