EDIT: Okay, I misunderstood the question at first, here's a real answer.
To read it easily and quickly, maybe record byte offsets at the beginning of the file for the first word of every letter (you may want to organize them alphabetically in the file) then from that letter have offsets for the second letter, then possibly the third letter, then check the words under that offset to see if any of them match. And what you may do to reduce file size is with the bite offset for example if the word was ARKANSAS and you had it under byte offset 43 under A and 28 under AR and 3 under ARK, instead of then writing the string ARKANSAS just put ANSAS, since you already have determined the first three letters by the location in the file based on the offsets.
Hopefully that was comprehensible.
|