how to read a txt file into an array

[nar0122]

I have a text file called A.txt that I am trying to read each character frequency from. The best way to show my problem is by example. Say the file reads "AbCAAb",

I have created an array of 256 memory positions to store each character frequency. So, first we encounter a capital A, which is the ascii value 65. So I go to position 65 in the array, and increment the count to 1. Then little b is 97, so I go to position 97 in the array...once I hit the next capital A, it should go to position 65 in the array and increment its counter to 2, then 3 for the next A. In the end, I should have an array consisting of zeros, where there were some characters that were not found in the file, and the frequencies of characters in the file that were found.

************************************************** ********************
THIS IS MY CODE THUS FAR
************************************************** ********************
//ignore the test points, they are for debugging purposes

int getFrequency(int arrayOfCharacters[])
{
    ifstream in;
    in.open("A.txt");
    cout<<"TEST POINT 2\n";
    for(int i = 0; i < 256; i++)
    {
        arrayOfCharacters[i] = 0;
    }
 
    cout<<"TEST POINT 3\n";
    int c = in.get();
    arrayOfCharacters[c]++;
 
 
    while(c!=EOF)
    {
    c = in.get();
    arrayOfCharacters[c]++;
 
    }
 
    for(int j = 0; j < 256; j++)
    {
        cout<<arrayOfCharacters[c];
    }
 
    in.close();
    return c;
 
}

[weaknessforcats]

Remember that arrays are 0 based. The 65th element for your A is array[64].

[whodgson]

So what succes have you had to date? What is your question?
Woops how do you read a text file into an array?....right?
Use the cin.getline () function after opening the file in read mode (r).

[nar0122]

So far I am getting the right results in the array aside from the last character in the array. Everything looks to be right, then I am getting a large number like 4677907 in the very last position in the array? I also need to know how to make the out put say, "There are (number of characters) (character)."

Example: File has "AAAaaa"

Output should read:

There are 3 A's.
There are 3 a's.

[weaknessforcats]

Are you using array[256] as your last element?

If you are, you need a 257 element array.

That would explain the garbage in the last element. Apparently my Post #2 wasn't clear enough.

[jfwfmt]

You are probably not checking for end of file correctly. Look at the documentation of .get() and its friends (google get())

/s/ Jim WIlliams

also note that 0-037, 0177 and 0200-0237 (octal) are not directly printable

[donbock]

Don't know about C++, but in C the char type can be either signed or unsigned. To be fully portable you should do something to insure that you never use a negative array index.

[donbock]

The following snippets show how to maximize portability by rooting out assumptions about implementation-dependent behavior.

#include <limits.h>        // Obtain UCHAR_MAX.
#define NCHARACTERS (UCHAR_MAX + 1)    // Don't assume 256.
...
int countingArray[NCHARACTERS];
...
int i, c;
...
for (i=0; i<NCHARACTERS; i++) {
    countingArray[i] = 0;
}
...
do {
    c = in.get();
    if (c != EOF)
    {
        i = ((unsigned char) c);    // Protect from compilers where char is signed.
        countingArray[i]++;
    }
} while (c != EOF);
...
for (i=0; i<NCHARACTERS; i++) {
    c = ((char) ((unsigned char) i));        // Recover printable character code (if you need it).
    ...
}

I'm not sure the casts on lines 16 and 22 are fully portable, but it seems reasonable to expect char and unsigned char to be interconvertible regardless of the integer encoding used by the compiler.

The double-cast on line 22 is probably excessive. On most (perhaps all) platforms it would be ok to cast directly from int to char.

[weaknessforcats]

I am going to offend someone, I just know it, but here goes:

1) The C++ header is <limits> and not <limits.h>. None of the standard C++ headers is guaranteed to be C-compatible and <limits> is a good example. All .h standard header are fo C. C++ standard headers have no extension and are in the std namespace.

2) In C++ you use the numeric_limits template for info on your type rather than the C sizeof.

3) There is no C cast in C++. The cast form (char) is compilable in C++ only as a backwards compatible feature. It doesn't work the same in C++ as in C. Using this in C++ will call conversion operator, if available. You are supposed to use the C++ cast forms.

4) The only time you need to cast in C++ is a) you are calling a relic C function that has a void* argument or some such or b) your C++ design is screwed up.

5) Prefer using vector to using arrays.

6) Prefer using the accumulates algorithm to writing loops.

7) Do not use macros in C++. There are several features in C++ that are direct maco replacements: a) templates, b) inline functions, c) enums, d) namespaces.