Man, you don't mess around!
One caveat for you about the epub, though, versus the PDF. If you downloaded it off of Google Books, the text may well be merely the rough OCR output from the PDF scan of the original book. That being the case, you may run into a few (or very, very many) errors in the text since, likely, the OCR text has not been gone over, cleaned up, corrected, and formatted by human editorship. And, I'm talking like huge chunks of text where virtually every word has at least one character that has been garbled, often rendering a mass of gibberish under which is barely recognizable what was written in the original document. At least, that has been my experience, from the times I've downloaded the text (*.epub, *.mobi) versions of old books from Google Books, and archive.org. However, ever since I got myself a Galaxy tablet a couple years ago, I stopped finding as much of a need for downloading old books in epub, or mobi, formats, because the tablet can hold sooooooo much more (in terms of memory (and thus, in terms of books)) than my old Kindle could hold, and so, the larger file sizes of the facsimile PDFs compared to the trifling sizes of the text files is no real issue. So, I just download the facsimile PDFs. Besides, in most cases, I really prefer the image of the original typeface, anyway, over the appearance of the sometimes drab, sterile-seeming system font.
On the other hand, though, as far as I know, as nerds are always making improvements in computer technology, OCR technology is no exception, and so, perhaps it's the case that the rough output of an OCR scan of more recent times is drastically much cleaner, and less error-laden, than one of, say, 5 years ago.