My inclination would still be to use DB_File or a relative to access a BTREE. One of the methods in the API is exactly equivalent to your search interface, but the implementation should be more performance-efficient. (And using their API takes less code.)
My inclination would still be to use DB_File or a relative to access a BTREE.
That's a good solution in some cases, but not always. The world is full of sorted text files,
and it doesn't always make sense to manufacture a much larger secondary file,
which must be maintained in parallel with the text file,
just to perform searches on the data.