no, you can do both. The trick is to use a other field for your file name.
Lets look at an example. Your file name is "10201155.pdf"
In the PDF you have define a title and an author. I use for this example the series field. You should create a custom column for your library to store the document number.
Go to the import dialog and put in (?P<series>.+) as regular expression, check that "read metadata for content rather then file name" is selected and apply the change. When you test it in the same window with file name "10201155.pdf" you will see the file name "10201155" in the field series. Now add one book for test (check before running that the metadata are in the file on the correct position).