Fix Microsoft Purview EDM Matching for Names and Emails
Question details
The user is experiencing inconsistent detection in Microsoft Purview Exact Data Match (EDM) when using FullName and EmailAddress as supporting columns, especially when data contains spaces, accented names, or the '@' character.
- Product
- Microsoft Purview
- Device & OS
- not provided
- Scenario
- Configuring EDM schemas and sensitive information types to accurately detect complex personal data using primary and secondary supporting elements.
- Observed behavior
- EDM fails to detect records when IdentityCard is the primary matching pattern and FullName or EmailAddress are supporting elements, though it successfully detects IdentityCard with Date of Birth. The system struggles with spaces and special characters.
Ensure you have the necessary compliance administrator permissions in the Microsoft Purview compliance portal and access to the EDM schema XML files and Sensitive Information Types (SITs) configurations.
Review EDM Normalization and Test Without Special Characters
Determine if normalization settings are causing the matching failure by temporarily testing data without spaces or special characters.
Microsoft Purview EDM relies heavily on how data is normalized before matching. If the base Sensitive Information Type (SIT) ignores certain characters, the Exact Data Match might fail to map the supporting elements properly.
While spaces and the '@' symbol are required in your real data, temporarily removing them in a test environment helps isolate whether the issue lies in the data format or the EDM schema logic.
Generate a small CSV file with synthetic test records. Ensure you include the IdentityCard, FullName, and EmailAddress columns.
In one version of the test dataset, remove all spaces from the 'FullName' field and the '@' symbol from the 'EmailAddress' field.
Use the EDM Upload Agent to hash and upload this sanitized test dataset to Microsoft Purview.
Test your EDM rule against documents containing the exact sanitized strings. If the match succeeds, you must adjust the normalization settings in your base SIT to appropriately handle spaces and the '@' character.
Configure Custom SITs for Supporting Elements
Ensure that the Sensitive Information Types used for supporting columns are properly configured to recognize the required data formats.
Handle Sensitive Documents Locally with WPS Office
While Microsoft Purview handles complex enterprise cloud compliance, occasionally you need a secure, local environment to review and manage sensitive datasets before uploading them. WPS Office provides a lightweight, highly compatible, and secure offline alternative to Microsoft Office.
- 1. Download WPS Office: Visit the official WPS Office website and download the free desktop application.
- 2. Install the application: Run the installer and follow the on-screen prompts to set up WPS Office on your computer.
- 3. Edit datasets securely: Open your CSV or XLSX datasets locally, apply necessary formatting for your EDM schemas, and save them securely.
Frequently Asked Questions
Does Microsoft Purview EDM support accented characters?
Yes, Microsoft Purview EDM supports accented characters, but the base Sensitive Information Type (SIT) and matching patterns must be specifically configured to recognize and process them without treating them as word breaks.
Why does EDM fail when using an email address as a supporting element?
This usually happens because the base SIT used to identify the email address normalizes the data incorrectly, or the '@' symbol causes tokenization issues that prevent the EDM engine from mapping it to the primary identifier.
Can I use FullName as a primary matching pattern in EDM?
Yes, you can use FullName as a primary matching pattern by creating a custom Sensitive Information Type (SIT) that uses a keyword list or regex, and mapping it correctly in your EDM schema.
How do I fix normalization issues in EDM?
You can fix normalization issues by reviewing your SIT configurations in the Purview portal. Ensure that the regex or keyword patterns account for spaces and special characters, and test your data using the exact data match test tools.
