hash with file instead of array

Question

Is it possible to use hash function but with File instead of Array, and it's gonna be saving the record in a file position and then search will fseek to that position, but I'm not sure how to open a file with let's say 1000 line, do you think that method could work?

• user40980: You are attempting to find a technical solution to a problem. The approach that you are using is likely to have some issues. What is the original problem that you are trying to solve?
• poly: Nothing really, I was just thinking if that is possible. for example can that be good enough to be used instead of a DB, I don't mean replace db, but for example if you want to use simple key with some data in a file.
• Lars Viklund: If you have an algorithm that works within a single slab of linear memory, you have something can be translated to a random-access file. You're going to have a hard time if you have variable-width records, handling collisions, rehashing, etc. You can implement a rudimentary system, but I would strongly recommend looking at a lightweight solution like SQLite3, leveldb (or possibly Berkley DB if you can accept copyleft).
• NoChance: You can even use a database for that - Depends on the problem.

Answer 1

Sure, if you have a hashing function that can neatly map your data. Two things to keep an eye on:

1. Does your OS/file system support sparse files? Can you seek beyond the end of the file and write a new record without intervening data?
2. How are you going to handle hash collisions? Will each record have some kind of header that contains a pointer to the next bucket in an overflow chain, or will you just overflow into the next available slot? If you just go to the next slot, how will you tell if it's available?

I'd recommend reading up on ISAM files, they're pretty close to what you're looking to do.

• poly: I'm using Linux and I think it does support it, and made also think about how to delete or replace data, that won't be I guess