I have built a data structure somewhat similar to a non clustered B+ tree index(on a field say K), over a data file with file offsets as my leaf node values. Now for any lookup, I need to read from a random point on the file. As I understand , most I/O methods on Java are optimized for bulk lookup. But since I already have a clustered index on another field, sorting on K is not an option. Is there any option in Java to optimize bulk reading of a file from random offsets ?
Thanks !!