Comparing a char

Question

So, I am trying to figure out the best/simplest way to do this. For my algorithms class we are supposed read in a string (containing up to 40 characters) from a file and use the first character of the string (data[1]...we are starting the array at 1 and wanting to use data[0] as something else later) as the number of rotations(up to 26) to rotate letters that follow (it's a Caesar cipher, basically).

An example of what we are trying to do is read in from a file something like : 2ABCD and output CDEF. I've definitely made attempts, but I am just not sure how to compare the first letter in the array char[] to see which number, up to 26, it is. This is how I had it implemented (not the entire code, just the part that I'm having issues with):

int rotation = 0;
char data[41];

for(int i = 0; i < 41; i++)
{
    data[i] = 0;
}

int j = 0;
while(!infile.eof())
{
    infile >> data[j+1];
    j++;
}

for(int i = 1; i < 27; i++)
{
    if( i == data[1])
    {
        rotation = i;
        cout << rotation;
    }
}

My output is always 0 for rotation. I'm sure the problem lies in the fact that I am trying to compare a char to a number and will probably have to convert to ascii? But I just wanted to ask and see if there was a better approach and get some pointers in the right direction, as I am pretty new to C++ syntax.

Thanks, as always.

[Read here about why you shouldn't use !infile.eof()](http://stackoverflow.com/questions/5605125/why-is-iostreameof-inside-a-loop-condition-considered-wrong) — Borgleader, Jan 14 '15 at 19:51
It sounds like you may be misreading the assignment. Was "`2ABCD`" given as an example? Where the character `'2'` means "Rotate two times"? Or did you decide that the character `'2'` will mean that? It sounds like a system that won't go above `'9'` very cleanly. — Drew Dormann, Jan 14 '15 at 19:56
If the 2 in your infile looks like a 2, the character value will not be 2. It will be 50, which is the ascii key for a 2. For a first try, if you want to get the value of a digit, just subtract '0'. — Hans Klünder, Jan 14 '15 at 19:57
That's a good point, Drew. Actually the assignment does give examples and they look much like that, however they do go above 9. I guess I missed that important detail so I suppose I will need to check if either the first or second element are integers. If they both are I will have to rotate a number of times greater than 9, depending on that specific number. — GiH, Jan 14 '15 at 20:03

R Sahu · Answer 1 · 2015-01-14T20:06:46.737

2

Instead of formatted input, use unformatted input. Use

data[j+1] = infile.get();

instead of

infile >> data[j+1];

Also, the comparison of i to data[1] needs to be different.

for(int i = 1; i < 27; i++)
{
   if( i == data[1]-'0')
            //      ^^^ need this to get the number 2 from the character '2'.
   {
      rotation = i;
      std::cout << "Rotation: " << rotation << std::endl;
   }
}

edited Jan 14 '15 at 20:06

answered Jan 14 '15 at 19:55

R Sahu

204,454
14
159
270

Thanks, I didn't know that was an option! Unfortunately, however, rotation = 0 still. – GiH Jan 14 '15 at 20:00

score 1 · Answer 2 · answered Jan 14 '15 at 20:17

You can do this using modulo math, since characters can be treated as numbers.
Let's assume only uppercase letters (which makes the concept easier to understand).

Given:

static const char letters[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZ";
const std::string original_text = "MY DOG EATS HOMEWORK";
std::string       encrypted_text;

The loop:

for (unsigned int i = 0; i < original_text.size(); ++i)
{

Let's convert the character in the string to a number:

  char c = original_text[i];
  unsigned int cypher_index = c - 'A';

The cypher_index now contains the alphabetic offset of the letter, e.g. 'A' has index of 0.
Next, we rotate the cypher_index by adding an offset and using modulo arithmetic to "circle around":

  cypher_index += (rotation_character - 'A'); // Add in the offset.
  cypher_index = cypher_index % sizeof(letters); // Wrap around.

Finally, the new, shifted, letter is created by looking up in the letters array and append to the encrypted string:

    encrypted_text += letters[cypher_index];
  } // End of for loop.

The modulo operation, using the % operator, is great for when a "wrap around" of indices is needed.

With some more arithmetic and arrays, the process can be expanded to handle all letters and also some symbols.

I think using a string that contains the letters `A...Z` doesn't make code clearer. Better subtract `'A'`, do calculations and add `'A'` back. Also, I think the divisor should be `strlen`, not `sizeof`. Or just `26` if you remove the list of the letters. — anatolyg, Jan 14 '15 at 21:08

score 0 · Answer 3 · answered Jan 14 '15 at 19:58

0

First of all you have to cast the data chars to int before comparing them, just put (int) before the element of the char array and you will be okay.

Second, keep in mind that the ASCII table doesn't start with letters. There are some funny symbols up until 60-so element. So when you make i to be equal to data[1] you are practically giving it a number way higher than 27 so the loop stops.

answered Jan 14 '15 at 19:58

Mackiavelli

432
2
12

Please don't put `(int)` in front of anything, these C-style casts are dangerous and should be avoided in C++. That said, why should you need to cast anything for comparison? – Ulrich Eckhardt Jan 14 '15 at 20:11
I disagree. To me, the comparison `f(x) = f(y)` is not clearer than `x = y` because it requires knowledge about `f`. Also, the two comparisons are not equivalent for all possible `f`. – Ulrich Eckhardt Jan 15 '15 at 18:58
LOL at tha tlast statement. – Mackiavelli Jan 15 '15 at 20:14
Do you want me to explain it further, so that you understand? – Ulrich Eckhardt Jan 17 '15 at 09:18

score 0 · Answer 4 · answered Jan 14 '15 at 20:04

0

The ASCII integer value of uppercase letters ranges from 65 to 90. In C and its descendents, you can just use 'A' through 'Z' in your for loop:

change

for(int i = 1; i < 27; i++)

to

for(int i = 'A'; i <= 'Z'; i++)

and you'll be comparing uppercase values. The statement

cout << rotation;

will print the ASCII values read from infile.

answered Jan 14 '15 at 20:04

Quip11

194
1
7

The operation is much easier performed by looping through the original text string rather than looping through the alphabet. – Thomas Matthews Jan 14 '15 at 20:18

Julian · Answer 5 · 2015-01-14T20:41:00.723

How much of the standard library are you permitted to use? Something like this would likely work better:

#include <iostream>
#include <string>
#include <sstream>

int main()
{
    int rotation = 0;
    std::string data;

    std::stringstream ss( "2ABCD" );
    ss >> rotation;
    ss >> data;

    for ( int i = 0; i < data.length(); i++ ) {
        data[i] += rotation;
    }

    // C++11
    // for ( auto& c : data ) {
    //     c += rotation;
    // }

    std::cout << data;
}

Live demo

I used a stringstream instead of a file stream for this example, so just replace ss with your infile. Also note that I didn't handle the wrap-around case (i.e., Z += 1 isn't going to give you A; you'll need to do some extra handling here), because I wanted to leave that to you :)

The reason your rotation is always 0 is because i is never == data[1]. ASCII character digits do not have the same underlying numeric value as their integer representations. For example, if data[1] is '5', it's integer value is actually 49. Hint: you'll need to know these values when handle the wrap-around case. Do a quick google for "ANSI character set" and you'll see all the different values.

Your determination of the rotation is also flawed in that you're only checking data[1]. What happens if you have a two-digit number, like 10?

It's not enough to just add the `rotation`: the ciphering should wrap-around, converting e.g. `Z` to `A` when `rotation = 1`, probably also leaving non-letters without change. — anatolyg, Jan 14 '15 at 21:12

Comparing a char

5 Answers5