Given large set of files, design a system to effeciently handle aquery like:
Find all files that have these words and doesn't have these words.

Followup optimization question:
Can you make it efficient for queries that involve words which are common across most files (not all) and not frequently queried.