Problem in program to count number of words, sentences and letters in C

Question

I needed to make a program to count the number of words, sentences and letters by getting an input from the user. The program works perfectly until the input i give is multi-lined. If the input is longer than the text that can fit in the terminal window, the program starts to ignore all full stops/question marks/exclamation marks. I don't know why, and I'd like some help. This doesn't happen if the text can fit in one line of the terminal window. I also printed every character when it's read by the program, but that also ignores all full stops/question marks/ exclamation marks. None of those characters get printed. For clarification, a sentence is just the number of full stops/question marks/ exclamation marks, number of words is just the number of spaces in the text plus 1. Here is my code:

#include <stdio.h>
#include <ctype.h> //for the isalpha() function
#include <cs50.h>  //for the get_string() function

int main(void)
{
    int sentences = 0, letters = 0;
    int words = 1;
    char character;
    string text = get_string("Enter Text: \n");

    char x = 0;

    while (text[x] != '\0')
    {
        character = text[x];
        switch (character)
        {
        case ' ':
            words++;
            break;
        case '.':
            sentences++;
            break;
        case '?':
            sentences++;
            break;
        case '!':
            sentences++;
            break;
        default:
            if (isalpha(character))
            {
                letters++;
            }
        }
        x++;
    }

    printf("\n");
    printf("WORDS: %d, LETTERS: %d, SENTENCES: %d\n", words, letters, sentences);
}

I'm fairly new to c, but I have around a year of experience in Python. Thank you for your time.

Please show the exact test input, expected result and actual result. — kaylum
Please change char x = 0; to int x = 0; so it does not overflow on long lines. — Weather Vane
Just because the terminal line overflows to / continues on the next line when it is full does not mean you have entered more then one "line". — Weather Vane

John Bode John Bode · Accepted Answer · 2020-12-22T12:59:49

I’m going to make a few suggestions.

First, don’t use get_string¹ (or scanf, or fgets). For a filter program like this, you don’t actually need to store the input in order to process it; use getchar (or fgetc) to read one character at a time and loop based on that:

int c;  // getchar returns int, not char
...
puts( "Enter Text:" );
while ( ( c = getchar() ) != EOF )
{
  // test c instead of text[x]
}

This approach will handle input of any length (such as if you redirect a file as your input), and it avoids the potential overflow issue Weather Vane identified in the comments. The downside is that you’ll have to manually signal EOF from the console for interactive input (using either Ctrl-z or Ctrl-d depending on your platform).

You can collapse some of your tests in your switch, such as

case '.' :     // Each of these cases "falls through"
case '!' :     // to the following case.
case '?' :
  words++;     // the end of a sentence is also the end of a word
  sentences++;
  break;

You’ll want to add cases to handle newlines and tabs:

case ' ' :
case '\n' :
case '\t' :
  words++;
  break;

except you don’t want to bump the words counter for repeating whitespace characters, or if the previous non-whitespace character was a punctuation character. So you’ll want an extra variable to track the class of the previously-read character:

enum {NONE, TEXT, PUNCT, WHITE} class = NONE;
...
while ( ( c = getchar() ) != EOF )
{
  switch( c )
  {
    case ' ' :
    case '\n' :
    case '\t' :
      if ( class == TEXT )
        words++;
      class = WHITE;
      break;
    
    case '.' :
    case '!' :
    case '?' :
      if ( class == TEXT )  // Don’t bump the word counter
        words++;            // if the previous character was
                            // was whitespace or .! ?
      if ( class != PUNCT ) // Don’t bump the sentence counter
        sentences++;        // for repeating punctuation
      
      class = PUNCT;
      break;
      ...
  }
}

There will still be weird corner cases where this won’t give a completely accurate count, but should be good enough for most input.

You should be able to figure out the rest from there.

^{The CS50 string handling and I/O routines like get_string are pretty slick, but they grossly misrepresent how C actually does things. The string typedef is especially egregious because what it aliases is not a string. Just be aware these tools will not be available outside the CS50 curriculum, so don’t become too reliant on them.}

Problem in program to count number of words, sentences and letters in C

1 Answers