word frequency of string counter is sometimes wrong - c

Question

I hope you can help me I worked on this code. The code works like this
user inputs a string for example "hey john, how are you john?
the program erases signs like "'?' , ',' '!' " etc.
the program writes a string after erasing the signs : "hey john how are you john?"
and the code outputs the frequency of each word:
hey : 1
john: 2
how : 1
are : 1
you : 1
but my code counts sometimes wrong. For example when I type "bye bye bye hello hello hello"
the output is :
bye : 3
hello : 1
My code does the john example right, but the bye bye... example wrong.
How do I have to change my code? Thank you
#include <stdio.h>
#include <string.h>
char words[80][80];
void clear_string(char *text);
int extract_and_count(char *source, int *count);
void clearArray(char array[]);
int indexInWords(char string[]);
void print(int countOfWords, int count[]);
int equals(char *string1, char *string2);
int main() {
char string[80];
int count[80];
printf("please enter your text: ");
scanf("%[^\n]s", string);
clear_string(string);
printf("%s\n", string);
int countOfWords = extract_and_count(string, count);
print(countOfWords, count);
return 0;
}
void clear_string(char *text){
int i = 0;
for(;i < strlen(text);++i){
if( text[i] == '.' || text[i] == ',' || text[i] == '!' || text[i] == '?'){
int k = i + 1;
for(; k < strlen(text);++k){
text[k-1] = text[k];
}
k = strlen(text) - 1;
text[k] = ' ';
}
}
}
int extract_and_count(char *source, int *count){
int wordCounter = 0;
char string[80];
int i = 0, k = 0;
clearArray(string);
for(; i < strlen(source);++i, ++k){
if(source[i] != ' '){
string[k] = source[i];
}else{
if(string[0] == '\0'){
break;
}
int index = indexInWords(string);
if(index == -1){
strcpy(words[wordCounter], string);
count[wordCounter] = 1;
wordCounter++;
}else{
count[index] += 1;
}
clearArray(string);
k = -1;
}
}
return wordCounter;
}
void clearArray(char array[]){
memset(array,0,strlen(array));
//array[0] = '\0';
}
int indexInWords(char string[]){
int i = 0;
for(;i < 80;++i){
if(equals(words[i], string) == 0){
return i;
}
}
return -1;
}
void print(int countOfWords, int count[]){
for(int i = 0;i < countOfWords; ++i){
printf("%s : %d\n",words[i], count[i]);
}
}
int equals(char string1[], char string2[]){
return strcmp(string1, string2);
}

Answer 1

The most significant problem I found was in extract_and_count() -- it doesn't count the last word as it only counts words followed by space. The bandaid is to check if string has anything in it after the loop, and if so, process it. Below is my rework for that fix and general style:
#include <stdio.h>
#include <string.h>
#include <stdbool.h>
void clear_string(char *text);
int extract_and_count(char *source, int count[]);
void clearArray(char array[]);
int indexInWords(char string[]);
void print(int countOfWords, int count[]);
bool equals(char *string1, char *string2);
#define BUFFER_SIZE (512)
#define MAX_WORD_COUNT (80)
#define MAX_WORD_SIZE (64)
char words[MAX_WORD_COUNT][MAX_WORD_SIZE];
int main() {
char string[BUFFER_SIZE];
int count[MAX_WORD_COUNT];
printf("Please enter your text: ");
while (fgets(string, BUFFER_SIZE, stdin) == NULL) {
printf("Please (re)enter your text: ");
}
clear_string(string);
int countOfWords = extract_and_count(string, count);
print(countOfWords, count);
return 0;
}
void clear_string(char *text) {
for (int i = 0; i < strlen(text); i++) {
if (text[i] == '.' || text[i] == ',' || text[i] == '!' || text[i] == '?' || text[i] == '\n') {
int length = strlen(text);
for (int k = i + 1; k < length; k++) {
text[k - 1] = text[k];
}
text[length - 1] = '\0';
i--;
}
}
}
int extract_and_count(char *source, int count[]) {
int wordCounter = 0;
char string[MAX_WORD_SIZE] = {'\0'};
for (int i = 0, k = 0; i < strlen(source); i++, k++) {
if (source[i] != ' ') {
string[k] = source[i];
} else {
if (string[0] == '\0') {
break;
}
int index = indexInWords(string);
if (index == -1) {
strcpy(words[wordCounter], string);
count[wordCounter] = 1;
wordCounter++;
} else {
count[index] += 1;
}
clearArray(string);
k = -1;
}
}
if (string[0] != '\0') {
int index = indexInWords(string);
if (index == -1) {
strcpy(words[wordCounter], string);
count[wordCounter] = 1;
wordCounter++;
} else {
count[index] += 1;
}
}
return wordCounter;
}
void clearArray(char array[]) {
memset(array, 0, strlen(array));
}
int indexInWords(char string[]) {
for (int i = 0; i < MAX_WORD_COUNT; i++) {
if (equals(words[i], string)) {
return i;
}
}
return -1;
}
void print(int countOfWords, int count[]) {
for (int i = 0; i < countOfWords; i++) {
printf("%s : %d\n", words[i], count[i]);
}
}
bool equals(char string1[], char string2[]) {
return strcmp(string1, string2) == 0;
}
The next most significant issue I see is you don't keep track of how many entries in words[][] are used, so indexInWords() could easily wander off making comparisons against uninitialized memory.

Answer 2

In extract_and_count you break out of the for-loop when you find 2 spaces. Also you did not check for the last word of source. Changed it to:
int extract_and_count(char *source, int *count){
int wordCounter = 0;
char string[80];
int i = 0, k = 0;
clearArray(string);
for(; i < strlen(source)+1;++i, ++k){
if(source[i] != ' ' && source[i] != 0){
string[k] = source[i];
}else{
if(string[0] != '\0'){
int index = indexInWords(string);
if(index == -1){
strcpy(words[wordCounter], string);
count[wordCounter] = 1;
wordCounter++;
}else{
count[index] += 1;
} }
clearArray(string);
k = -1;
}
}
return wordCounter;
}