If you want to use the fact that the data are sorted, you can just read the file once in order to build an array of integers that are the character offsets of the start of each line. (If all the lines are the same length, you don't even need to do this!) Then use file seek (direct access) and binary search to find the start of a range, reading lines until you get to the end of the range.
In this manner you eliminate the overhead of building a huge data structure (a file offset is only 6 or 8 bytes). But you will be doing log_2(n) seeks to find each range start, which could be too slow if you need to do many many lookups.
