A problem I see is that for each of your lines you are going to perform each of your substitutions. That is, with N names and M pairs, you are doing N * M substitutions.

Is that really necessary? Is there more about the data, be it the names or the pairs that can be used to speed up the process? Do you really need the N * M substitutions, or are you just replacing words with words?

Abigail