Question

Writing a program that uses a linked list stack to convert an equation in infix notation to postfix notation. The stack portion of the program is its own class and is in its own header file, and is implemented correctly (able to compile and run the test main provided by my professor). I’m currently working on the actual function to convert an infix string to a postfix string in an inherited class. I have the following code, which I think has done the shunting yard algorithm incorrectly somewhere since I get odd output.

template <class myType>
void infixToPostfix<myType>::convertToPostfix()
{
    linkedStack<char> newS; //creating a new char stack
    string output; //creating the postfix output string

    for(int i = 0; i < infx.length(); i++) //cycle through the entire string
    {
        if (isgraph(infx[i])) //skip spaces
        {
            if (isalpha(infx[i])) //if the char is a letter (caps or uncaps)
            {
                output += infx[i]; //append it to output
            }
            else
            {
                if (newS.isEmptyStack()) newS.push(infx[i]);
                else if (precedence(infx[i], newS.top()) //check if the current char has higher precedence than the top of the stack, or if the stack is empty
                {
                    newS.push(infx[i]); //push the current char onto the stack
                }
                else if (infx[i] == ')') //check if the current char is a closing paren
                {
                    for(;;)
                    {
                        if (newS.isEmptyStack()) break;
                        if (newS.top() != '(') //check if the top of the stack isn't an open paren
                        {
                            output += newS.top(); //append the top of the stack to the output
                            newS.pop(); //pop the top of the stack off
                        }
                        else //the top of the stack is a (
                        {
                            newS.pop(); //pop the ( off the stack
                            break; //break out of the for loop
                        }
                    }
                }
                else //the current char doesn't have higher precedence than the top of the stack
                {
                    output += newS.top(); //append to the top of the stack to output
                    newS.pop(); //pop off the top of the stack
                    newS.push(infx[i]); //put the current char onto the top of the stack
                }
            }
        }
    }

    while (!newS.isEmptyStack()) //not sure if this works, assuming we're at the end of the line at this point, and if there's anything on the stack we need to append it to the output
    {
        output += newS.top();
        newS.pop();
    }

    pfx = output; //setting pfx to the output (pfx is the class variable for the postfix output)

}

My function to show the postfix string is as follows

template <class myType>
void infixToPostfix<myType>::showPostfix()
{
    cout << "Postfix Expression: " << pfx << endl;
}

I get the following output when I run the program.

silverbox@silverbox-VirtualBox:~/Dropbox/CS202/ass13$ ./a.out testinput1.txt
Infix Expression: A + B - C
-ostfix Expression: AB+C
Infix Expression: A + B * C
*+stfix Expression: ABC
Infix Expression: A * B + C / D
/+stfix Expression: AB*CD
Infix Expression: (A + B) * C
*+(tfix Expression: AB)C
Infix Expression: A * (B + C) / D
/+(tfix Expression: A*BC)D
Infix Expression: (A + B) * (C - D)
-(+(fix Expression: AB)*CD)
Infix Expression: A * (B + C / D)
/+(tfix Expression: A*BCD)
Infix Expression: A + ((B + C) * (E - F) - G) / (H - I)
-(--(+( Expression: A+(BC)*EF)G)/HI)
silverbox@silverbox-VirtualBox:~/Dropbox/CS202/ass13$ 

I honestly don’t understand why what looks like odd bits of the postfix expression get pushed over onto the string in my cout. Any hints/help?

Edit: My output is as follows after making the change suggested by ymett. I’m now attempting to figure out where I went wrong in attempting to handle parentheses.

silverbox@silverbox-VirtualBox:~/Dropbox/CS202/ass13$ ./a.out testinput1.txt
Infix Expression: A + B - C
Postfix Expression: AB+C-
Infix Expression: A + B * C
Postfix Expression: ABC*+
Infix Expression: A * B + C / D
Postfix Expression: AB*CD/+
Infix Expression: (A + B) * C
Postfix Expression: AB)C*+(
Infix Expression: A * (B + C) / D
Postfix Expression: A*BC)D/+(
Infix Expression: (A + B) * (C - D)
Postfix Expression: AB)*CD)-(+(
Infix Expression: A * (B + C / D)
Postfix Expression: A*BCD)/+(
Infix Expression: A + ((B + C) * (E - F) - G) / (H - I)
Postfix Expression: A+(BC)*EF)G)/HI)-(--(+(
silverbox@silverbox-VirtualBox:~/Dropbox/CS202/ass13$ 

Answer 1

You haven’t shown how infx is filled, but it would appear that it has a carriage return (CR, ‘\r’, 0x0d, 13) at the end.

Your condition infx[i] != ' ' should be replaced by a condition that checks for any blank character, ie !isblank(infx[i]). Better yet, instead of checking for characters you don’t want, check for characters you do want; if you don’t have a list use isgraph(infx[i]) (any character with a visible representation, ie not a control characters or a space).

The condition infx[i] >= 65 && infx[i] <= 122 isn’t good either. Firstly you should use character literals rather than numbers, ie infx[i] >= 'A' && infx[i] <= 'z'. Secondly you have included characters between 'Z' and 'a' which are not letters, so it should be (infx[i] >= 'A' && infx[i] <= 'Z') || (infx[i] >= 'a' && infx[i] <= 'z'). Also you are assuming letters are contiguous in the character set which is true for ASCII but not for all character sets (although perhaps we don’t have to worry about that). Better use the tools the language gives you, and write isalpha(infx[i]).