Converting if statement to conditional statement

Question

I am trying to convert an IF statement as below

SET RESULT = IF(C>(X-2), (A+B), C) //IF C< X-2 return A+B otherwise return C

to a conditional statement as below

SET RESULT = C>(X-2)?(A+B):C

I have wrote a code which scan the whole string and look of occurrences of IF , ( and , . My algorithm is not working when there more than 1 IF statement such as below

SET RESULT = IF(C>(X-2), IF(P>C,2,3),C)

Here is the code...

        string data = "SET RESULT=IF(C>(X-2), (A+B), C)";
        string output = string.Empty;

        int indexofIF = data.IndexOf("IF");
        int obCount = 0;
        int cbCount = 0;
        if (indexofIF > -1)
        {
            string script = data.Substring(indexOfIF, (data.Length-indexOfIF-1))
            for(int index=0; index<script.Length; index++)
            {
                int obIndex = data.IndexOf('(', index);
                if(obIndex>-1)
                    obCount++;
                int cbIndex = data.IndexOf(')', index);
                if(cbIndex>-1)
                    cbCount++;

                if(obCount==cbCount)//Found the end of If statement
                {
                    string tempData = data.Substring(0, index);
                    int count = tempData.Count(f => f == ',');//Get the number of occurences of ','
                    if (count == 2)//There are only 2 commas
                    {
                        int firstIndex = tempData.IndexOf(',');
                        int lastIndex = tempData.LastIndexOf(',');
                        string condtion = tempData.Substring(3, (firstIndex - 4));
                        string trueCond = tempData.Substring(firstIndex + 1, (lastIndex - firstIndex - 1));
                        string falseCond = tempData.Substring(lastIndex + 1, (index - lastIndex - 1));
                        output = condtion + "?" + trueCond + ":" + falseCond;

                    }
                    else //More than 2 commas
                    {

                    }
                }

            }

        }

I am not sure if this will work for complex scenarios. Is there any better other way of doing this? Perhaps using regex or any other string replacement operation..

Answer 1

You are on the right track.

I would:

• Read each character into a small buffer
• evaluate the buffer to look for the word "IF"
• set a flag so you know you are in an IF statement
• clear the buffer
• evaluate for parenthesis
• keep track of opening/closing of parenthesis
• perform any additional logic (such as comma parsing)
• continue until the end of the string

Perhaps using regex or any other string replacement operation..

(IF.*?(?=IF|$)) breaks out multiple IF statements but it doesn't handle the additional parsing and requires that the input data follow a strict structure. Keeping track of parenthesis is tricky/impossible (see comments on this answer) and more easily handled by a state machine.

Personally, I like the control/flexibility of a character-by-character parse. Tools like Code Mirror use this approach for parsing source code.

You may find that a regular expression is useful once you have successfully extracted a portion of the string (eg to deconstruct a formula like A + B > C into its parts).