Counting word occurrences in a file C - c

Question

Welcome everybody. I am new to Stackoverflow, I code in C for some time.
I have run to a problem writing a program counting word occurrences in a text file. I need to have an output telling what word occurred how many times. Here is the source code:
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
int new_words=0;
int nwords=0;
typedef struct element{
char word[30];
int how_many;
} element;
int is_word_new(element ** dictionary, char * string)
{
for (int i =0; i<new_words; i++)
{
if (strcmp(string, dictionary[i]->word)==0)
return 0;
}
return 1;
}
int which_word(element ** dictionary, char * string)
{
for (int i =0; i<new_words; i++)
{
if (strcmp(string, dictionary[i]->word)==0)
return i;
}
return 0;
}
int main()
{
FILE * fp;
char word[30];
fp=fopen("input.txt", "r");
if (fp==NULL)
{
printf("FILE ERROR");
return 0;
}
while(!feof(fp))
{
fscanf(fp, "%s",word);
nwords++;
}
nwords--;
rewind(fp);
struct element * dictionary = (element*)malloc(sizeof(element)*nwords);
for (int i =0; i<nwords; i ++)
{
fscanf(fp, "%s", word);
if( is_word_new(&dictionary, word) )
{
strcpy(dictionary[new_words].word, word);
//dictionary[new_words].word= word;
dictionary[new_words].how_many=1;
new_words++;
}
else
dictionary[which_word(&dictionary, word)].how_many++;
word[0]='\0';
}
printf("\n\nFinal dictionary\n with %d words", new_words);
for (int i =0; i<new_words; i++)
{
printf("%s %d \n", dictionary[i].word, dictionary[i].how_many);
}
free(dictionary);
fclose(fp);
return 0;
}
the idea is that i first count how many words are in the text (which somehow is always greater by one than in fact). The function is_word_new checks if a newly read word is already in the dictionary. which_word() tells which word was found
However I get a segmentation fault running this program.
When I used the line which is commented // dictionary[i].word=word the program behaved as if there was only "word" in the dictionary.
Please give me hints where am I doing this stuff wrong

Answer 1

Must read question: Why is “while ( !feof (file) )” always wrong? Thanks to Jonathan Leffler's comment.
Please check my comments in the code below. I got you a start up for when the words are appearing once. I am letting the rest of the job for you, so that we can share the fun, but you can of course ask.
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
int new_words = 0;
int nwords = 0;
typedef struct element {
char word[30];
int how_many;
} element;
// no need to pass double pointer
int is_word_new(element* dictionary, char * string) {
int i;
for (i = 0; i < new_words; i++) {
printf("|%s|, |%s|\n", string, dictionary[i].word);
if (strcmp(string, dictionary[i].word) == 0)
return 0;
printf("i=%d\n",i);
}
return 1;
}
int which_word(element ** dictionary, char * string) {
int i;
for (i = 0; i < new_words; i++) {
if (strcmp(string, dictionary[i]->word) == 0)
return i;
}
return 0;
}
int main() {
FILE * fp;
char word[30];
fp = fopen("test.txt", "r");
if (fp == NULL) {
printf("FILE ERROR");
return 0;
}
printf("file read\n");
int read_counter;
while (!feof(fp)) {
read_counter = fscanf(fp, "%s", word);
// increment only if we really read something
if(read_counter >= 0)
nwords++;
}
// this is wrong, remove it
//nwords--;
rewind(fp);
printf("nwords = %d\n", nwords);
// do not cast what malloc returns. Also struct is not needed.
element * dictionary = malloc(sizeof (element) * nwords);
int i;
for (i = 0; i < nwords; i++) {
fscanf(fp, "%s", word);
printf("read |%s|\n", word);
if (is_word_new(dictionary, word)) {
strcpy(dictionary[new_words].word, word);
//dictionary[new_words].word= word;
dictionary[new_words].how_many = 1;
new_words++;
} else {
printf("bhka\n");
dictionary[which_word(&dictionary, word)].how_many++;
}
//word[0] = '\0';
}
printf("\n\nFinal dictionary\n with %d words", new_words);
for (i = 0; i < new_words; i++) {
printf("%s %d \n", dictionary[i].word, dictionary[i].how_many);
}
free(dictionary);
fclose(fp);
return 0;
}
Here is the test.txt I used:
sam klouvi george dit epfl
ok
end