VB.Net/C#-自定义函数上的拆分字符串

Question

I'm working on a custom mathematical expression calculator, but I'm having problems at parsing nested conditional expression like this one:

IIF("M"="M",(IIF(100 < 50,(IIF(2 > 0.45,2,1)),(IIF(2 > 0.45,4,3)))),(IIF(100 < 46,(IIF(2 > 0.45,2,1)),(IIF(2 >0.45,4,3)))))

What I'd like to do is to split the IIF function by commas in order to get its parameters:

Dim condition = "M"="M"
Dim truePart = (IIF(100 < 50,(IIF([2 > 0.45,2,1)),(IIF(2 >0.45,4,3))))
Dim falsePart = (IIF(100 < 46,(IIF(2 > 0.45,2,1)),(IIF(2 >0.45,4,3)))))

At the moment I'm using Regex to parse single IIF function by getting what is inside the parentheses and the split it by commas:

\((.*?)\)

Obviously that doesn't work with such expression since it will stop at the first closing parentheses, therefore I thought about using this to get all the other characters:

\((.*?)\).*

But now I'm not sure how to split it, since using commas is not an option anymore.

Answer 1

The answer from theory is that regular expressions are not capable to do what you requested because they "cannot count". However, you need to count.

The practise says that .NET regular expressions are no regular expressions but stack machines. With a group (?<Group>.*) you in fact add an entry to a stack of that group. With (?<-Group>) , you can remove an entry from that stack. You can also test whether the stack is empty.

Out of curiosity, I gave it a try and I believe that

[\\(,]([^\\(\\)]|(?<Par>\\()|(?<-Par>\\)))*(?(Par)---|[,\\)])

should work, where --- is used as an escape sequence. If you understand that "regular expression" right away, then I think you are good to go. In all other cases, I would rather recommend you to write a parser manually. Otherwise, you are not going to understand your code 5min after you have tested it.