(admin: I would recommend a title change for this, perhaps something like "How to populate newly exposed columns for large volume of cases").
We have had to do this a couple of times for our application. There are a couple of things to consider: technical approach and timing of execution.
There are some different technical approaches you could take: write Pega activities to populate the data and commit the items OR write database scripts.
I would recommend that if possible you create a clone of your production system to do a dry run of your data population so that you can make sure it works as expected and so that you can get an idea of the time that it will take to execute completely on production.