Sparse Files
I have keyed my answers below to your questions. HTH, Cheers, Dave Charles Sullivan Yes, to do this you should edit the meta information which is part of every track. Your tracks may not now...
Tony Lawrence
I also doubt that any inventory system would be done this way. I would expect that a hash function would be applied to the inventory numbers unless they were already large and disparate enough to get any benefit from sparse storage. If they were entirely sequential, you'd want to modulo them down to an offset from 0, and there wouldn't be any sparseness unless there were large gaps. If you had numbers like you show above, you'd probably need to put a hash function on them to spread them out to indirect (double, triple) blocks or they wouldn't cause sparseness again.
It's not necessarily easy to design a good hash for a given set of data. If the set is fairly small, it might be better to try to condense the keys down so that everything would be in early blocks, because there is a small penalty for double and triple indirects.
We used to care about this kind of thing a lot. With the performance of modern systems and cheap storage, you have to be storing a tremendous amount of data that you need quick access to before you'd even start thinking about this stuff now.
Linux groups from Newsgroups