Exact data match sensitive info type, or EDM SIT extends and builds on the sensitivity info type enabling you to protect specific data instances. You can use EDM to protect confidential information relating to specific clients, employees, patients, or the like. Exact data match functionality is available under an E5 subscription and is able to identify or protect up to 100 million unique values.

You start by defining the shape or format of the unique data in the form of a schema. This can be done by example, as in uploading a file of sample data or by manually configuring the schema. A sample file must have a column header row, where the column names must only contain alphanumeric characters, so no spaces or underscores. When you upload the file, any data that matches a sensitive info type will be identified. Alternatively, you can define a data format manually by selecting a sensitive info type. The sensitive info types can be either predefined or ones you've created yourself. 

I'll quickly go through the process of adding one column called client id that is formatted as a US bank account. Each EDM sensitive info type needs at least one primary column, which should contain unique values. I'll go with the default column settings and leave my detection confidence level as high. Click next and submit.

Once the EDM data schema has been defined and saved, you need to upload instances of the data you want to protect. Sensitive data is uploaded using the EdmUploadAgent command line tool. Most customers will use the commercial version, but there are government and department of defense versions. The EdmUploadAgent tool hashes the sensitive data as it's uploaded with a built-in salt value, or you can provide your own salt value. Sensitive data, such as patient IDs are seldom static, so you can use the EdmUploadAgent to upload and refresh to a data store up to twice a day.

If you created your EDM schema through the wizard as we have, you must use that schema file with the EdmUploadAgent tool.