Text dictionary like transformation in Google DLP

Question

I would like the data to be masked, but it was possible to understand how many people studied at UNIVERSITY_1.

What de-identification transformation can I use to accomplish such information\text masking?

Input:

{
    "students": [
        {
            "name": "John Smith",
            "university": "University of Pennsylvania"
        },
        {
            "formattedName": "Mike Miller",
            "university": "Harvard University"
        },
        {
            "formattedName": "Elon Musk",
            "university": "University of Pennsylvania"
        }
    ]
}

Output:

{
    "students": [
        {
            "name": "John Smith",
            "university": "UNIVERSITY_1"
        },
        {
            "formattedName": "Mike Miller",
            "university": "UNIVERSITY_2"
        },
        {
            "formattedName": "Elon Musk",
            "university": "UNIVERSITY_1"
        }
    ]
}

Answer 1

You could use crypto hashing: https://cloud.google.com/dlp/docs/deidentify-sensitive-data#cryptohashconfig

Answer 2

You could create a custom infotype with a single dictionary item (1 per college) and do a replace with infotype transform - how many schools are in your data set?