I am trying to search through the file counts the no. of ‘C’ reserved words encountered using binary search

Question

I am writing a program in c that searches through source code files and counts the number of ‘C’ reserved words encountered. But the reserved word is printed only if the reserved word entered is first word. And it counts total number of strings not total no of reserved word used. Can someone help me on this. My code is so messy please don't mind that.

#include <stdio.h>
#include <string.h>
#include <ctype.h>
#include <stdlib.h>
#define KEYMAX 32


FILE *fp ;
char data[1024];
struct keyword
{
    char word[10];
    int occur;
};
int i = 0, j = 0, pos;
char str[100], unit[20], ch;
int stored[1024];
char delimiters[] = " \t\n\v\f\r";  /* possible space delimiters */
char *token;
struct keyword key[32] = {"auto", 0, "break", 0, "case", 0,
                          "char", 0, "const", 0, "continue", 0,
                          "default", 0, "do", 0, "double", 0,
                          "else", 0, "enum", 0, "extern", 0,
                          "float", 0, "for", 0, "goto", 0,
                          "if", 0, "int", 0, "long", 0,
                          "register", 0, "return", 0, "short", 0,
                          "signed", 0, "sizeof", 0, "static", 0,
                          "struct", 0, "switch", 0, "typedef", 0,
                          "union", 0, "unsigned", 0, "void", 0,
                          "volatile", 0, "while", 0,};

int main()
{
    takeinput();
    system("CLS");
    theresult();
    // processresult();
    // ctoken();
return (0);

}

int takeinput()    // function to write in the file
{

  printf( "**********Welcome*************" ) ;
    fp = fopen("test.c", "w") ;   // Open file in write mode.
    if ( fp == NULL )
    {
        printf( "Could not open file test.c" ) ;  // Prints the statement if the file is not able to open.
        return 1;
    }
    printf( "\nPlease enter some text from keyboard to write in the file test.c \n\t" ) ;
    // getting input from user
    while ( strlen ( gets( data ) ) > 0 )
    {
        // writing in the file
        fputs(data, fp) ;           // Writes to file
        fputs("\n", fp) ;
    }

    // closing the file
    fclose(fp) ;
    return 0;
}

int theresult()
{

   fp = fopen("test.c", "r"); // read mode

   if (fp == NULL)
   {
      perror("Error while opening the file.\n");   // Prints the statement if the file is not able to open.
      return 1;
   }
   printf("The contents of test.c file are:\n");

 // To covert the ch into str
   int i= 0;
  //  printf("-----this is from ch----\n"); (Just for reference)
   while((ch = fgetc(fp)) != EOF)
   {
    str[i]=ch;
    i++;

   //  printf("%c",ch);   prints character

   }

   printf("%s",str);

   // printf("\n----This is from token-----\n");   (just for reference)
   for (token = strtok(str, delimiters); token != NULL;
         token = strtok(NULL, delimiters)) /* 'for loop' conditional part */
        /* prints token one per line */
       // puts(token);      // prints token

    for (i = 0; i < strlen(str); i++)
    {
        while (i < strlen(str) && str[i] != ' ' && isalpha(str[i]))
        {
            unit[j++] = tolower(str[i++]);
        }
        if (j != 0)
        {
            unit[j] = '\0';
            pos = binarysearch(unit, key);
            j = 0;
            if (pos != -1)
            {
               key[pos].occur++;
            }
        }
    }
    printf("***********************\n   Keyword\tCount\n***********************\n");
    for (i = 0; i < KEYMAX; i++)
    {
        if (key[i].occur)
        {
            printf("  %s\t  %d\n", key[i].word, key[i].occur);       // Prints the reserved keyword and its occurance
        }
    }

    fclose(fp);
   return (0);
}



int binarysearch(char *word, struct keyword key[])
{
    int low, high, mid;

    low = 0;
    high = KEYMAX - 1;
    while (low <= high)
    {
        mid = (low + high) / 2;
        if (strcmp(word, key[mid].word) < 0)
        {
            high = mid - 1;
        }
        else if (strcmp(word, key[mid].word) > 0)
        {
            low = mid + 1;
        }
        else
        {
            return mid;
        }
    }
    return -1;
}

The string entered is: if i break please re-join it. float float

---

Keyword Count

---

if            1 
break     1
Float     2

Answer 1

The error is in function theresult. In the for loop for tokenizing the input you process and search a word from the whole input str instead of the word token returned from strtok. You would not have to check for space (' ') after tokenizing because space is part of the delimiters.

It seems to work for me after changing the loop to:

   for (token = strtok(str, delimiters); token != NULL;
         token = strtok(NULL, delimiters)) /* 'for loop' conditional part */
        /* prints token one per line */
       // puts(token);      // prints token

        for (i = 0; i < strlen(token); i++)
        {
            while (i < strlen(token) && token[i] != ' ' && isalpha(token[i]))
            {
                unit[j++] = tolower(token[i++]);
            }
            if (j != 0)
            {
                unit[j] = '\0';
                pos = binarysearch(unit, key);
                j = 0;
                if (pos != -1)
                {
                   key[pos].occur++;
                }
            }
        }

The output is

The contents of test.c file are:
if i break please re-join it. float float
***********************
   Keyword      Count
***********************
  break   1
  float   2
  if      1

Additional remarks:

Instead of reading the whole input file into str I suggest to read and process the input line by line using fgets and a loop.

If you want the user to enter the input text you could process the input lines directly instead of writing them to a file "test.c" first and then reading the file.

The while loop after tokenizing will cut off everything from the first non-alpha character. Maybe you implemented this because it did not work due to the original error. When the token is "re-join", it will search for "re". You should check if this is what you want and change the while loop if necessary.