User statistic scheme with large amount of data

Consult the user statistics scheme with large data volume
For example, an app has 500W users.
Generate 500 txt files per day, each file corresponding to a user's action record, including (user name, IP, Access page, Mac, length of visit and other fields)
Now it is necessary to filter the content of 500W files, what is the best solution?

What I can think of now is to import this data into the MySQL database and then count the filters, but the amount of data is too large to know if it works. And how to quickly import a database is also a problem.



------Solution--------------------
Learn about the number of users who have not yet met this level of order

