Thread: alphabetically sort textfile

Heya,

I need to alphabetically sort a textfile and store the ordered result in another textfile.

How can this be solved?

Code:

#define LENGTH 255

int main()
{
    FILE *filein;
    FILE *fileout;
    char puffer[LENGTH];
    filein=fopen("namein.txt","r");
    fileout=fopen("nameout.txt", "a");

while(fgets(puffer,LENGTH,filein))
{
        fputs(puffer,fileout);
}

thought of the following solutions:
1. idea: store the data in some array and then just use a regular sort algorithm on it.
1. issue: I don't know the amount of entries in advance, thus the size of the array is unknown. I mean I would have to once read through the full textfile to count lines and then create the array based on that information, or is there a better approach?

2. idea: use a second puffer to read through the textfile and sort on the fly by comparing puffer1/puffer2 and storing the very "smallest" in the new textfile
2. issue: sort algorithm needs me to to "delete" or flag entries of the first textfile once they were put into into the second one, and I'm not quite sure if that's even possible with the "while fgets puffer" approach.

could you suggest me some ideas on how to best solve this issue? :/ maybe with tries or something even, I dunno :X

thanks in advance

U guys don't know puffer?



Anyways, thanks for the info, I will go with idea #1 then.

I can't think of a solution of the mentioned issue though: how can I set a dynamic size of my array? I mean, my array can't know how many elements it will have to store unless I once count the lines in my textfile (and I want to prevent having to once loop through it). is there a solution to that?

Basically the idea is, that the amount of lines my textfile has should not be limited to a specific, set number. Only the linelength is pregiven with 200 characters.

Thx in advance ^^

DID YOU SAY YOUR LINE LENGTHS WERE PREGIVEN AT 200 CHARS?

A-L-W-A-Y-S 200 chars? I'll bet you could find a relationship between a file with x number of lines:

x = file_size in bytes / 200 chars, where each char is one byte long.

(maybe plus a small fudge factor - strictly for "engineering safety") < lol >

You'll have a few bytes for the file header and probably uncounted newlines at the end of each of the lines of text, but that should give you a good approximation of the number of lines in the file.

Test that out, see what you can discover.

fseek() and ftell() are not standard C (library)?

Originally Posted by coffee_cup

I can't think of a solution of the mentioned issue though: how can I set a dynamic size of my array?

You allocate some initial memory and when you need more you call realloc().
11.3 Reallocating Memory Blocks

Bye, Andreas

Thanks all for the input ^^

Originally Posted by AndiPersti

You allocate some initial memory and when you need more you call realloc().
11.3 Reallocating Memory Blocks

Bye, Andreas

Nice, that should do the trick, thx a ton

Originally Posted by stahta01

For small (less than 2 GB files), fseek() and ftell() works good on getting file size.
(Technically the file size needs to fit in signed int variable to work.)

A signed long int, rather. There are a few issues with using fseek and ftell to retrieve the size of a file. Firstly, the value returned by ftell only gives you the number of bytes from the beginning of the file for binary streams, so if you're using a text stream the value needn't represent the number of bytes that the stream has progressed from the beginning. Also, streams with a binary mapping can have multiple null characters appended at the end of them, so seeking to the end of a binary stream may not necessarily give you the position of the end of the file.

...seeking to the end of a binary stream may not necessarily give you the position of the end of the file.

This is very true. It is, in fact, undefined behavior.

Setting the file position indicator to end-of-file, as with fseek(file, 0, SEEK_END), has undefined behavior for a binary stream (because of possible trailing null characters) or for any stream with state-dependent encoding that does not assuredly end in the initial shift state.

- 7.19.3, footnote 228

It might not be standard C, but the OS seems like a good place to get the file size.