C# Infix to Postfix convertor not giving the right answer if few case

Question

I made a Infix to Postfix convertor, and thought it worked, but when I went back and showed it to my teacher one of the examples he tested turns out wrong. :|

I would be grateful if someone could help me on this matter, and let me know what's wrong.

I skipped the parts about the buttons for inputting numbers and only posted the rest of it,

    private void button20_Click(object sender, EventArgs e)
    {
        try
        {
            string infix = textBox1.Text;
            infixTopostfix obj = new infixTopostfix(infix);
            textBox1.Text = string.Empty;
            Console.WriteLine("{0} is the Postfix of  {1}", obj.createPrifex(), infix);

        }
        catch (Exception e1)
        {
            Console.WriteLine(e1.ToString());
        }
    }

The part above is for redirecting a console.writeline to my text box and its the button that does all the works and gives the final result.

Here is the main class:

class infixTopostfix
{
    public infixTopostfix(string strTemp)
    {
        strInput = strTemp;
    }

    private int isOperand(char chrTemp)
    {
        char[] op = new char[6] { '*', '/', '+', '-', '^', '(' };
        foreach (char chr in op)
            if (chr == chrTemp)
            {
                return 1;
            }
        return 0;
    }
    private int isOperator(char chrTemp)
    {
        char[] op = new char[5] { '*', '/', '+', '-', '^' };
        foreach (char chr in op)
            if (chr == chrTemp)
            {
                return 1;
            }
        return 0;
    }

    private string strResualt = null;
    private string strInput = null;
    public string createPrifex()
    {
        int intCheck = 0;
        //int intStackCount = 0;
        object objStck = null;
        for (int intNextToken = 0; intNextToken <= strInput.Length - 1; intNextToken++)
        {
            intCheck = isOperand(strInput[intNextToken]);
            if (intCheck == 1)
                stkOperatore.Push(strInput[intNextToken]);
            else
                if (strInput[intNextToken] == ')')
                {
                    int c = stkOperatore.Count;
                    for (int intStackCount = 0; intStackCount <= c - 1; intStackCount++)
                    {
                        objStck = stkOperatore.Pop();
                        intCheck = isOperator(char.Parse(objStck.ToString()));
                        if (intCheck == 1)
                        {
                            strResualt += objStck.ToString();
                        }
                    }
                }
                else
                    strResualt += strInput[intNextToken];

        }//end of for(int intNextToken...)
        int intCount = stkOperatore.Count;
        if (intCount > 0)
        {
            int c = stkOperatore.Count;
            for (int intStackCount = 0; intStackCount <= c - 1; intStackCount++)
            {
                objStck = stkOperatore.Pop();
                intCheck = isOperator(char.Parse(objStck.ToString()));
                if (intCheck == 1)
                {
                    strResualt += Convert.ToString(objStck);
                }
            }
        }

        return strResualt;
    }

    private System.Collections.Stack stkOperatore = new System.Collections.Stack();

}

}

Here is the input that fails:

A^B^(C-D/(E+F))-(G+H)^L*Z+Y

the result of this code which is incorrect:

ABCDEF+/-^^GH+-LZY+*^

The correct result:

ABCDEF+/-^^GH+L^Z*-Y+

Answer 1

When converting infix notation to postfix notation, AKA reverse polish notation, one must take into account operator precedence and operator associativity. These are usually implemented as a table and looked up in the table based on the operator character.

e.g.
Java
C#
C++

As I do not see any mention of precedence or associativity in your code; so I would say your code does not have a bug but an invalid algorithm.

The classic algorithm to convert infix to postfix is the shunting yard algorithm by Edsger Dijkstra and is probably the means closest to what you are trying to implement.

I would suggest that you work with pen an paper to understand the shunting algorithm then implement the algorithm using many test cases that progressively use more combinations of the operators.

The best explanation I know of is Shunting Yard Algorithm

If you Google for C# shunting yard you will find many implementations, just make sure if you use one it is correct. Many people like to post examples of this but often have a mistake. Verify they work against many test cases before you waste time with them.