Thread: Letter Frequency counting.

Hi,

I'm still fairly new at C and have been asked to build a program that will count a number of specified letters to be read from a text file. This is the second stage of the task, the first being to write a program that reads a text file, removes all the spaces and converts the message to upper case, saving it in a new output text file.

Code:

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <ctype.h>

void main()
{

	FILE  *fpin, *fpout, *fprec;                        // declares two file streams, in and out
	int tl, f, i=0;
	int nl=0;
	char ch, let=0;                                    // ch is variable to be used in conversion
	char input[80], output[80], record[80];					// declared arrays to store file paths
	printf("Input input file name:  \n");
	gets(input);                                // input file location
	printf("Input output file name: \n");
	gets(output);								// output file location
	printf("Input record file name: \n");
	gets(record);								// record file location
		
	
	if((fpin = fopen(input,"r")) == NULL)		// confirms that file has been opened
	{      
		printf("Unable to open input file");
		exit(0);
	}

	if((fpout = fopen(output,"w")) == NULL)		// confirms that file has been created
	{    
		printf("Unable to open output file\n");
		exit(0);
	}

	tl=0;

	while((ch=getc(fpin))!=EOF)                // read input file until the end of the file
	{
	    if((ch>='a' && ch<='z') | (ch>='A' && ch<='Z')) // ensures that only alphabetic characters are transferred
		{   
		putc(toupper(ch), fpout);               //function to change all the characters to upper case
	    tl++;
		}
		
	}

	fclose(fpin);								// close input file
	fclose(fpout);

	fopen(output,"r");

	if((fprec = fopen(record,"w")) == NULL)		// confirms that file has been created
	{    
		printf("Unable to open record file\n");
		exit(0);
	}

	printf("Total number of letters in the file: %d\n", tl);	

	printf("Which letter would you like the frequency of?\n");
	scanf("%c", let);

This is the area I am having difficulty with. My thinking is that if I can get ch into normal string format I will be able to use a for loop to go through each character and determine whether or not it is the character that the user has input.

Code:

	while((ch=getc(fpout))!=EOF)                // read input file until the end of the file
	{
		char c = tolower(ch);
		if(let=c)
			{
				nl++;
			}
	}

	f=(nl/tl)*100;
		
	printf("%d", f);

	putc(f, fprec);

	fclose(fpout);
	fclose(fprec); 
}

Unfortunately however, I either don't know how to do this, or as a result of having spent the last 6 hours banging my head on this thing I have simpley forgotten. If anyone can fill me in or see a more logical way of doing this I'm open to all suggestions.

Thanks for your time.

James

Code:

if(let=c)

That's an assignment, not a comparison.

Do you have warnings turned on, turned up and did you fix them? I'd say no...

Code:

	
        scanf("%c", &let);                       // <-- need ampersand
        while((ch=getc(fpout))!=EOF)                // read input file until the end of the file
        {
                char c = tolower(ch);                   // why go though the trouble to convert data saved to upper, only to reconvert to lower?
                if(let==c)                          // <-- single '=' is assignment, not comparison

If you're only looking for one specific character all you need to do in your old program is add a couple of lines...

1) You need to get the user's character... that's easy...
2) in your file loop, catch the character before you filter it with if (ch>='a' ... and increment a counter.
3) to report either write the count to the file or display it on the screen... This file had 232 occurances of the letter E.

Should be no more than one or two extra lines...

Thank you very much for the fast responses. About the void thing, from the moment my lecturer told us what void did it always seemed like a silly idea, but she insists on the main function being void. =\

It's now time for some light revision, so I'll come back to this tomorrow when my brain is fresh. What's been said has made sense and taken on board, thank you all again.

To back up hk_mp5kpdw and Tater, there is an overview of the right and wrong ways to declare main here. At the bottom of that page, it links to a nice article that explains the problems of void main in more detail. Show them to your teacher, just be respectful and diplomatic about it.

Originally Posted by CommonTater

2) in your file loop, catch the character before you filter it with if (ch>='a' ... and increment a counter.

Still having a pretty rough time with this. I'm also using fgetc now, as this says that the pointer will be a character instead of an integer. Judging from what you've said this will be wrong, but I'm really not sure how to "catch" the letter without using an if statement. This is what I have:

Code:

	while((ch=fgetc(fpout))!=EOF)                
	{	
		if(ch>='a' && ch==&let)
				{
					nl++;
				}
	}

nl is not incremented. I have tried every variation of that if statement that I can think of, but to no avail. It's done when I get that nl int to increment.

Cheers.

Is there some reason you're trying to read characters from a file stream opened for output?

Code:

while((ch=getc(fpin))!=EOF)                // read input file until the end of the file
  {

     // check for the user's counted character here!

     if((ch>='a' && ch<='z') | (ch>='A' && ch<='Z')) // ensures that only alphabetic characters are transferred
  {

Yes, I have to have this all in one program. So Input is the first file, I take the data from that, remove spaces and convert to caps, stick it in output. Then take the message in output, count the total letters, count the number of user specified letters and store the result of the equation: f=(number of letters/total number of letters)*100 into the record file. The next stage of the task is to convert it all into seperate functions, which shouldn't be too hard at all.

I left out:

Code:

if((fpout = fopen(output,"r")) == NULL)		
	{
		printf("Unable to open output file\n");
		exit(0);
	}

C) nicely compartmentalized functions. If I have to choose between those options C generally wins.

Absolutely. Once a few options have coalesced in my head, I invariably pick the one that seems to offer me the most testable (compartmentalized) code.

For example, if you need to read a file, do a bunch of stuff and then write to a file, forget the "bunch of stuff" at first. Start by creating a simple "test" file in a text editor, then write code to open that file and print out what it contains to the screen. Once that works, write code to create a new file and write simple data to it. Once that works, save the source file somewhere safe, and move forward by modifying a copy of it (you may need the original when all hell breaks loose later)! Then in your code combine the two by making your code read one file and write its data to a second file. The work on the "bunch of stuff".

I deffinitely get what you're both talking about, it's logical to keep each function short and sweet, but I'm afraid this damn comparision between ch and let is still doing my head in.