Multidimensional arrays returning.

[eXeCuTeR]

Well, actually I'm trying to create a program that receives strings from a file and replaces a few words from some kind of a source (another file) and swap between words that are in the source of the words.
These words are shown this way in the file:

elephant fat
mouse thin
smart stupid

and in the original file there are strings, such as:

This guy is really smart!

So now, I'm trying to receive the data from the strings-file and swap words that are shown in the source file and create a new sentence by swapping these words.
e.g:
original file:

You are smart

After compiling:

You are stupid.

Here's what I've got so far:

#include <stdio.h>
#include <string.h>

#define LENGTH 50

char *newSubstr(char *str, int index);
char **explode(char *str, char seperator);

int main(void)
{
    FILE *handle, *db;
    char *filename, str[LENGTH];
    
    printf("File: ");
    fgets(filename, 16, stdin);
    
    handle = fopen(filename, "rt");
    db = fopen("database.txt", "rt");
    
    while(fgets(str, LENGTH, handle) != NULL)
    {
                     for(int i = 0; i < strlen(str); i++)
                     {
                            if(str[i] == ' ' && i + 1 < strlen(str))
                            {
                                      char search[LENGTH], **results;
                                       while(fgets(search, LENGTH, db) != NULL)
                                       {
                                                  results = explode(search, ' ');
                                                  if(strcmp(results[0], newSubstr(str, i+1)) == 0) // if equals, swap the string with results[1]
                                                                        printf("Match! &#37;s - %s", results[0], results[1]);                                                               
                                       }
                            }
                     }                          
    }
    
 getchar();
 return 0;   
}

char *newSubstr(char *str, int index)
{
 int i = index, j= 0;
 char *newstr = "";
 
 for(; i < strlen(str); i++)
 {
       if(str[i] == ' ')
                 break;
       newstr[j++] = str[i]; 
 }
                
 return newstr;
}


char **explode(char *str, char seperator) // explodes ONLY 2 strings that are seperated by a character
{
     int i = 0, index = 0;
     for(; i < strlen(str); i++)
           if(str[i] == seperator)
                     break;
     
     char **newexplode;
     if(i != 0 && i != strlen(str) - 1)
     {
        int index_0 = 0, index_1 = 0;
        int j = 0;
        for(; j < i; j++)
           newexplode[0][index_0++] = str[j];                          
        for(; i < strlen(str); i++)
           newexplode[1][index_1++] = str[i];                           
     }
     
     return newexplode;
}

I'm pretty sure (actually, I know) that there is a problem with the return values of `explode` and `newSubstr`...I dunno, look at this code for example:

char *newSubstr(char *str, int index)
{
 int i = index, j = 0;
 char newstr[20];
 
 for(; i < strlen(str); i++)
 {
       if(str[i] == ' ')
                 break;
       newstr[j++] = str[i]; 
 }
                
 return newstr;
}

int main(void)
{
    char *str = newSubstr("Hello world", 6);
	printf("%s", str);
    getchar();
 return 0;   
}

Really basic but still, `str` contains total garbage, and I don't know why...



It's pretty basic what I've done here so please help me ASAP (these are just pointers and return values, come on )

[Elysia]

>>char **newexplode;
>>newexplode[0][index_0++] = str[j];
You never allocate any memory for newexplode.

[anon]

It seems that you are nowhere allocating memory for the exploded strings?
And anyway, there is the strtok function that might be suitable for splitting strings.

In newSubstr you are returning a pointer to a local array. The problem with local arrays is that they go out of scope and the pointer is not valid once the function returns. (In addition newstr doesn't appear to get 0-terminated.)
Usually it is just easier to pass the buffer that receives output as an argument. Think about strcpy: it doesn't allocate any new memory, it copies to the provided buffer.

[eXeCuTeR]

> Elysia, Why do I need to allocate memory for `newexplode`?
> anon, Why I'm supposed to 0-terminate newstr?

Usually it is just easier to pass the buffer that receives output as an argument. Think about strcpy: it doesn't allocate any new memory, it copies to the provided buffer.

What do you mean?

[anon]

>
> anon, Why I'm supposed to 0-terminate newstr?

What do you mean?

Strings in C must be terminated with the '\0' character. This is the way how any function such as strlen or printf can actually know where the string ends.

As to how newSubstr could be implemented:

#include <stdio.h>
#include <string.h>

void newSubstr(char* output, const char *str, int index)
{
    int i = index, j = 0;

    /*strlen is a bit too expensive to call each time in a loop*/
    int length = strlen(str);

    for(; i < length; i++)
    {
        if(str[i] == ' ')
            break;
        output[j++] = str[i];
    }
    /*null-terminate the output string*/
    output[j] = '\0';
}

int main(void)
{
    /*declare the array that receives the substring and pass it to newSubstr*/
    char str[20];
    newSubstr(str, "Hello world", 6);
	printf("%s", str);
    getchar();
 return 0;
}

[anon]

As to why you need to allocate memory for newexplode: if you don't make it to point to some memory, the pointer will point to random garbage and you'd end up trying to write to a random place in memory that your program might not own.

However, it is not easy to allocate (and free) two-dimensional arrays.

Instead you might simply modify the original string, taking advantage of how strings are represented in C, and return a pointer that points in the middle of the original string where the second word begins.

original string:
'm','o','u','s','e',' ','t','h','i','n','\0'
 ^
 |
str

after splitting:
'm','o','u','s','e','\0','t','h','i','n','\0'
 ^                        ^
 |                        |
str                    start of
                      second word

[eXeCuTeR]

Strings in C must be terminated with the '\0' character. This is the way how any function such as strlen or printf can actually know where the string ends.

As to how newSubstr could be implemented:

#include <stdio.h>
#include <string.h>

void newSubstr(char* output, const char *str, int index)
{
    int i = index, j = 0;

    /*strlen is a bit too expensive to call each time in a loop*/
    int length = strlen(str);

    for(; i < length; i++)
    {
        if(str[i] == ' ')
            break;
        output[j++] = str[i];
    }
    /*null-terminate the output string*/
    output[j] = '\0';
}

int main(void)
{
    /*declare the array that receives the substring and pass it to newSubstr*/
    char str[20];
    newSubstr(str, "Hello world", 6);
	printf("&#37;s", str);
    getchar();
 return 0;
}

As I've heard, C automatically 0-terminates arrays, doesn't it?
Anyways,
I've got 1 questions:
Why do local arrays go out of scope if they are returned as pointers?
As for explode...I've tried to implement it this way:

void explode(char *str, char seperator, char **match) // explodes ONLY 2 strings that are seperated by a character
{
     char first[20], second[20];
     int i = 0, index = 0;
     for(; i < strlen(str); i++)
           if(str[i] == seperator)
                     break;
     
     if(i != 0 && i != strlen(str) - 1)
     {
        int index_0 = 0, index_1 = 0;
        int j = 0;
        for(; j < i; j++)
           first[index_0++] = str[j];                          
        for(; i < strlen(str); i++)
           second[index_1++] = str[i];    
           
        strcpy(match[0], first);
        strcpy(match[1], second);                       
     }
}

But it's not working properly.
Why?

[anon]

As I've heard, C automatically 0-terminates arrays, doesn't it?

Yes, library functions do that. But if you manipulate individual characters yourself, you have to take care of terminating strings.

Anyways,
I've got 1 questions:
Why do local arrays go out of scope if they are returned as pointers?

That's because they are local. You would be returning a pointer to something that doesn't exist after the function returns.

As for explode...I've tried to implement it this way:
...snip...

But it's not working properly.
Why?

For one thing, you are still not terminating the new strings.

For another thing, which can't be seen from this code: have you allocated any memory for match in the calling code (and how)?

Anyway, you might take a look at my previous post to see how the same thing could be done without allocating any new strings.

[eXeCuTeR]

Yes, library functions do that. But if you manipulate individual characters yourself, you have to take care of terminating strings.

That's because they are local. You would be returning a pointer to something that doesn't exist after the function returns.


For one thing, you are still not terminating the new strings.

For another thing, which can't be seen from this code: have you allocated any memory for match in the calling code (and how)?

Anyway, you might take a look at my previous post to see how the same thing could be done without allocating any new strings.

Oh right, forgot to 0-terminate.
---
I haven't allocated any memory to any array, I have just called it this way, which seems totally cool for me:

    char *newarr[2];
    explode("Hello world", ' ', newarr);
    printf("%s [ ~ ] %s", newarr[0], newarr[1]);

I would like to hear the way to do this by allocating memory and the other way.

[anon]

It may look cool to you but it is still nothing more than an array of two pointers to garbage.

It should look something like this:

int main()
{
    char *newarr[2];
    newarr[0] = malloc(20);
    newarr[1] = malloc(20);
    explode("Hello world", ' ', newarr);
    printf("%s [ ~ ] %s", newarr[0], newarr[1]);
    free(newarr[0]);
    free(newarr[1]);
    return 0;
}

However, if you just modified the original string in place (as my ASCII art shows) it might be implemented like this (you can roll your own if you are not allowed to use library functions).

#include <stdio.h>
#include <string.h>
#include <stdlib.h>

char* explode(char* str, const char* separator)
{
    char* second = strtok(str, separator);
    if (second != NULL) {
        second = strtok(NULL, separator);
    }
    return second;
}

int main()
{
    char str[] = "Hello world";
    char* second = explode(str, " ");
    printf("%s [ ~ ] %s", str, second);
    return 0;
}

[eXeCuTeR]

It works good now with the memory allocation.
Why can't I do this thing:

char *arr[2] = (char *)malloc(40);

I should allocate memory to each pointers(or ARRAYS) I use (in case he is not pointing to any variable)? I'm assuming that yes since pointers that are not initialized point to a random memory address.

[eXeCuTeR]

If anyone is interested...

#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#define LENGTH 80

void explode(char *str, char seperator, char **match, int index1, int index2) // explodes ONLY 2 words that are seperated by a character
{
     int i = 0, index = 0;
     for(; i < strlen(str); i++)
           if(str[i] == seperator)
                     break;
     
        int index_1 = 0, index_2 = 0;
        int j = 0;
        for(; j < i; j++)
           match[index1][index_1++] = str[j];  
        i++;                        
        for(; i < strlen(str); i++)
        {
           if(str[i] == '\n')
                     break;
           match[index2][index_2++] = str[i]; 
        }
               
        match[index1][index_1] = '\0';
        match[index2][index_2] = '\0';                    
}

void newSubstr(char* output, char *str, int index)
{
    int i = index, j = 0;

    /*strlen is a bit too expensive to call each time in a loop*/
    int length = strlen(str);

    for(; i < length; i++)
    {
        if(str[i] == ' ' || str[i] == '\n')
            break;
        output[j++] = str[i];
    }
    /*null-terminate the output string*/
    output[j] = '\0';
}

int main(void)
{
    FILE *handle, *db;
    handle = fopen("hi.txt", "r");
    db = fopen("database.txt", "r");
    
    int len = 0;
    char newDBstring[LENGTH], *results[LENGTH/2], str[LENGTH];
    
    for(int i = 0; i < LENGTH/2; i++)
            results[i] = (char *)malloc(LENGTH/2);
             
    int index1 = 0, index2 = 1;
    while(fgets(newDBstring, LENGTH, db) != NULL)
    {
                             explode(newDBstring, ' ', results, index1, index2);
                             printf("&#37;s,%s\n", results[index1], results[index2]);
                             index1 = index1 + 2, index2 = index2 + 2;
                             len++;
    }
    
    while(fgets(str, 40, handle) != NULL)
    {
                     printf("\n%s\n", str);
                     for(int i = 0; i < strlen(str); i++)
                     {
                             if(str[i] == ' ' || i == 0)
                             {
                                       int index = 0;
                                       char *output = (char *)malloc(LENGTH);
                                       if(i == 0)
                                            newSubstr(output, str, i);
                                       else
                                           newSubstr(output, str, i+1);
                                       printf("Output: %s\n", output);
                                       for(int k = 0; k < len*2; k = k+2)
                                       {
                                               if(strcmp(output, results[k]) == 0)
                                                                     printf("MATCH! %s = %s -> %s\n", output, results[k], results[k+1]);        
                                       } 
                                       free(output);
                             }        
                     }                
    }
    fclose(handle);
    fclose(db);
    for(int i = 0; i < len; i++)
            free(results[i]);
    
    getchar();   
    return 0;
}

[vlrk]

The Output for your program is

smart,hard
pass,fail

You are smart

Output:You
Output:are
Output:smart
MATCH! smart = smart -> hard

Is this is the same output which is as per design.

I thought you must get like

If it is

You are smart

You are stupid

where database.txt is

smart stupid

and hi.txt is

You are smart

[C_ntua]

It works good now with the memory allocation.
Why can't I do this thing:

char *arr[2] = (char *)malloc(40);

I should allocate memory to each pointers(or ARRAYS) I use (in case he is not pointing to any variable)? I'm assuming that yes since pointers that are not initialized point to a random memory address.

I think you would write it like:

char (*arr)[2] = (char (*)[10])malloc(40);

so you typecast in an appropriate thing. The other will not work, since *arr[2] isn't type of char*

[anon]

char (*arr)[2] = (char (*)[10])malloc(40);

I get an "initialization from incompatible pointer type" warning. And when I try it, it indeed doesn't do what I want.

Since it is an array of 2 pointers, I believe you can't get away without two memory allocations (or assigning two pointers):

char* arr[2] = {malloc(20), malloc(20)};

const char* str = "Hello\0world";
const char* arr[2] = { str, str + 6 };