Thread: Making unique integer ID# from a string.

I've got a bunch of text files, with unique names. I want to make an array in my program of which each cell represents one of those filenames.
For instance
array[file_ID] holds information about one of the tekstfiles.

How can I turn a tekstfile name which is a string, into a unique file_ID number?
Eventually i'm going to use this array as an array of counters. Each file needs its own counter.

If the files already had such a number, eg, the files are named "123.txt", "4.txt", "6667.txt", I would say this idea is okay. If you are just going to assign arbitrary numbers to them for the sake of the program, I would say it is a bad idea, the reason being that you will need to look up the name somehow anyway (so why not just use the name in the first place? If you have to do a lookup to get the unique_id based on the filename in order to find out this "counter" (another number?) for that file, you will need two arrays...and you will have to search one of them for a value).

This is probably a good place to use a linked list, which if you haven't used one, now is the time to learn (unless you don't intend to ever write anything but very simplistic C programs, and you may not be able to do what you want that way in the end*). Then you could have the information stored in a struct, and search the list using the actual filename.

It is unclear what you mean by counter in this context.

* Sorry to break that to you but it's true; you may need to do a few days of work learning about linked lists before you proceed. Or keep hacking away and wind up in the same place. Trust me You may not need an actual linked list, but you will at least need an array of pointers to a more complex datatype, either a string or a struct. Start reading and you'll figure it out.

Originally Posted by Adak

What about using a key and encrypting the filename?

The ASCII chart comes to mind, using two digits for every letter or number in the filename.

The numbers you choose could be different than ASCII for a very mild type of security or convenience.

Neat idea, but since you probably want at least the first three characters in order to distinguish one file from another, you will have to assign an array of 262626 elements (no upper or lower case). Straight ASCII values would be 127127127 elements (so Abcsomething.txt would be array[41098099], and abcsomething.txt array[97098099]).

Beyond three letters, think just plain stupid, unless you use a more complex formula to compress and encrypt.

Originally Posted by Adak

What about using a key and encrypting the filename?

The ASCII chart comes to mind, using two digits for every letter or number in the filename.

The numbers you choose could be different than ASCII for a very mild type of security or convenience.

What you're describing is a hash function, albeit an extremely expensive one. I'd stay away from cryptographic hashes and use something simpler, but the idea of hashing is probably the way to go.

Originally Posted by brewbuck

What you're describing is a hash function, albeit an extremely expensive one. I'd stay away from cryptographic hashes and use something simpler, but the idea of hashing is probably the way to go.

In which case w/r/t our astrophysics student (the OP) perhaps C is not the appropriate language to be learning.

I thought of a way around this problem now, so I don't need it anymore.
I don't want to make it too complicated because i'm the only one that's using the program to analyze some data, and I shouldn't spend too much time on the program itself.
This can be one for when I get enough spare time .

Originally Posted by Ayreon

This can be one for when I get enough spare time .

Kind of how I might feel about astrophysics

Keep in mind about what I was trying to saying about learning the language -- there are probably better choices than C if you just need to write basic utility programs for your own purposes.

Within our education there is a programming course which everyone must follow. They used C, and that's why I'm using C, it's the only language I know anything about and I hear it is a good thing to understand C. Learning other languages is apparently easier with a C background (that's what I hear people say). Anyway, with help from this forum I'm actually starting to enjoy programming a little , so for the time being I'm gonna stick with C.

Originally Posted by Ayreon

Within our education there is a programming course which everyone must follow. They used C, and that's why I'm using C, it's the only language I know anything about and I hear it is a good thing to understand C. Learning other languages is apparently easier with a C background (that's what I hear people say). Anyway, with help from this forum I'm actually starting to enjoy programming a little , so for the time being I'm gonna stick with C.

Won't knock you for that! That a computer language will be easier to understand if you already know another one is something I would assume is generally true, regardless of whether C is one of them or on what side of the equation it is. So I'd thought I'd mention it since this could be a real "speak now or hold your peace moment" for you. The only other language I'm competent with is perl, which I still use when I think it will do what I want, well enough -- eg. perl has hashes, so

Originally Posted by perl

my %files;
$files{Abc.txt}="anything 17";
$files{xyz.jpg}=789+x;

which is a few more lines in C

Originally Posted by Adak

@MK, since all letters "map" or correspond to 2 digit numbers, there is no need for what you describe. Of course, the length of the numbers would be twice the length of the filename char's, originally.

Adak, it will be exactly what I describe because you cannot just guess the first letter and subtract 42 or something. Using ASCII values you would need three digits for each letter, so the highest real value, if "zzz", would be array[122122122].

Originally Posted by MK27

this could be a real "speak now or hold your peace moment" for you.

I'm not sure I understand what you mean by that, I'm not marrying C . I am open to suggestions for using other languages(The one you just mentioned looks handy), but not right now, because I've already started with C and I wouldn't want to recreate my program in another language, which I would have to learn from scratch.

But Isn't it true that most languages are based on C, and that that's why it is easier to learn other languages if you know C?

Originally Posted by Ayreon

But Isn't it true that most languages are based on C, and that that's why it is easier to learn other languages if you know C?

I think that most languages are based on "lower level" concerns, but C is more transparently close to those concerns than most other languages -- which creates this illusion.

That's also why something like perl is easier to use for certain things -- it is much less concerned with low level transparency.