Hash Table in C (find the frequency of every word) - c

Question

I want to create a hash table for an exercise I have to send in my University.
The program will open a number of files, break each file's content to <<words>> (tokens) and it will save each <<word>> in a hash table with the frequency of each <<word>>.
In case the word is already in the hash table , the program will increase the word's frequency.
At the end the program will print the words and it's frequencies accordingly.
Also the frequencies should be printed from the highest word frequency to the lowest.
The comparison of the <<words>> will ignore upper and lower case letters.
For example if a file contains : one two three four Two Three Four THREE FOUR FoUr
It should print:
four 4
three 3
two 2
one 1
The professor gave us a template that we should complete but I'm really confused on what to do with the insert_ht() and clear_ht() functions as well as the compare one.
Here is the code :
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <ctype.h>
#define HTABLE_SIZ 1001
#define MAX_LINE_SIZ 1024
/* Hash Table */
typedef struct node* link;
struct node { char *token; int freq; link next; };
link htable[HTABLE_SIZ] = { NULL }; /* Table of lists (#buckets) */
int size = 0; /* Size (number of elements) of hash table */
unsigned int hash (char *tok );
void insert_ht (char *data);
void clear_ht ( );
void print_ht ( );
void Process(FILE *fp);
int main(int argc, char *argv[])
{
int i;
FILE *fp;
for (i=1; i < argc; i++)
{
fp = fopen(argv[i],"r");
if (NULL == fp)
{
fprintf(stderr,"Problem opening file: %s\n",argv[i]);
continue;
}
Process(fp);
fclose(fp);
}
print_ht();
clear_ht();
return 0;
}
void Process(FILE *fp)
{
const char *seperators = " ?!'\";,.:+-*&%(){}[]<>\\\t\n";
char line[MAX_LINE_SIZ];
char *s;
while((fgets(line,MAX_LINE_SIZ, fp)) != NULL)
{
for (s=strtok(line,seperators); s; s=strtok(NULL,seperators))
insert_ht(s);
}
}
/* Hash Function */
unsigned int hash(char *tok)
{
unsigned int hv = 0;
while (*tok)
hv = (hv << 4) | toupper(*tok++);
return hv % HTABLE_SIZ;
}
void insert_ht(char *token)
{
……………………………………………
}
void clear_ht()
{
……………………………………………
}
int compare(const void *elem1, const void *elem2)
{
……………………………………………
}
void print_ht()
{
int i, j=0;
link l, *vector = (link*) malloc(sizeof(link)*size);
for (i=0; i < HTABLE_SIZ; i++)
for (l=htable[i]; l; l=l->next)
vector[j++] = l;
qsort(vector,size,sizeof(link),compare);
for (i=0; i < size; i++)
printf("%-50s\t%7d\n",vector[i]->token,vector[i]->freq);
free(vector);
}

Answer 1

I'll answer you in a new post because it's hard to be exhaustive in comments.
1. Malloc
Why would I need to use malloc then ? Shouldn't i write directly to the htable? (on the insert_ht() funtion)
You need to use malloc because you declare a char pointer in struct (char *token). The thing is that you never initialize the pointer to anything, and as far you don't know the size of the token, you need to malloc every token. But, as you use strdup(token), you don't need to malloc token because strdup does. So don't forget to free every token in order to avoid memory leaks.
2. Segfault
I can't test you code, but it seems like the following line causes the segmentation fault :
list = htable[hashval]->token
Indeed, you try to access token while htable[hashval] is NULL, and to assign a char * to a link type (list).
You need to loop with this :
for(list = htable[hashval]; list != NULL; list = list->next) { ... }
3. Notes
if (x=1) should be if(x==1).
Don't malloc new_list if you don't need to.
Because new_list if used when htable[hashval] is NULL, new_list->next = htable[hashval]; will set new_list->next to NULL.
You should use the -Wall option in gcc (for warnings) and you may use valgrind to understand your segmentation faults. In this case, use gcc with debug mode (-g).

Answer 2

Double and Final edit : Ι found the solution. Apparently for some reason my compare function was wrong.
I still haven't figured out why but here is the correct one, hopefully someone else will find this post helpful!
int compare(const void *elem1, const void *elem2)
{
return (*(link*)elem2)->freq - (*(link*)elem1)->freq;
}
Edit: deleted old answer . Found the correct way I think but I have another problem right now.
The compare function doesn't work correctly. My printf is fine but it doesnt sort them with the frequiencies. I want them to be sorted from the highest to lowest .
In this example: the file contains -> one two three four Two Three Four THREE FOUR FoUr
And I get:
two 2
one 1
four 4
three 3
While I should be getting :
four 4
three 3
two 2
one 1
Here is the code. Feel free to help!
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <ctype.h>
#define HTABLE_SIZ 1001
#define MAX_LINE_SIZ 1024
/* Hash Table */
typedef struct node* link;
struct node { char *token; int freq; link next; };
link htable[HTABLE_SIZ] = { NULL }; /* Table of lists (#buckets) */
int size = 0; /* Size (number of elements) of hash table */
unsigned int hash (char *tok );
void insert_ht (char *data);
void clear_ht ( );
void print_ht ( );
void Process(FILE *fp);
int main(int argc, char *argv[])
{
int i;
FILE *fp;
printf("prin tin for \n");
for (i=1; i < argc; i++)
{
printf("prin tin fopen \n");
fp = fopen(argv[i],"r");
if (NULL == fp)
{
fprintf(stderr,"Problem opening file: %s\n",argv[i]);
continue;
}
printf("prin tin process \n");
Process(fp);
fclose(fp);
}
print_ht();
//clear_ht();
return 0;
}
void Process(FILE *fp)
{
const char *seperators = " ?!'\";,.:+-*&%(){}[]<>\\\t\n";
char line[MAX_LINE_SIZ];
char *s;
while((fgets(line,MAX_LINE_SIZ, fp)) != NULL)
{
for (s=strtok(line,seperators); s; s=strtok(NULL,seperators)){
printf("prin tin insert %s \n",s);
insert_ht(s);
}
}
}
/* Hash Function */
unsigned int hash(char *tok)
{
printf("bike stin hash \n");
unsigned int hv = 0;
while (*tok)
hv = (hv << 4) | toupper(*tok++);
printf("VGAINEIIIIIIIIIIIIII %d \n",hv);
return hv % HTABLE_SIZ;
}
void insert_ht(char *token)
{
printf("bike stin insert %s \n",token);
unsigned int hashval = hash(token);
if (htable[hashval]==NULL){
printf("mesa stin prwti if %u %s \n",hashval,token);
//token = strdup(token);
htable[hashval] = malloc(sizeof(token));
htable[hashval]->token = token ;
htable[hashval]->freq = 1;
size++;
}else {
htable[hashval]->freq++;
}
printf("ta evale epitixws \n");
}
int compare(const void *elem1, const void *elem2)
{
const struct node *p1 = elem1;
const struct node *p2 = elem2;
if ( p1->freq < p2->freq)
return -1;
else if (p1->freq > p2->freq)
return 1;
else
return 0;
}
void print_ht()
{
int i, j=0;
link l, *vector = (link*) malloc(sizeof(link)*size);
for (i=0; i < HTABLE_SIZ; i++)
for (l=htable[i]; l; l=l->next)
vector[j++] = l;
qsort(vector,size,sizeof(link),compare);
for (i=0; i < size; i++)
printf("%-50s\t%7d\n",vector[i]->token,vector[i]->freq);
free(vector);
}

Answer 3

Sorry for my bad english.
I think that :
insert(char *token) takes a word of the file and puts into the hash table. In brief, if the word exists in the hash table, you just have to increment its frequencie. Otherwise, you need to create another node and put the frequencie to 1, then ad it to the array. At the end, you will have one entry for each unique word.
compare(const void *elem1, const void *elem2) will be used by qsort. It returns 0 if elem1 = elem2, a negative number if elem1 < elem2 and a number > 0 if elem1 > elem2. By passing compare to qsort, you allow qsort to sort you array according to your own criteria.
clear_ht() may set all the values of the array to NULL, in order to restart another count ?