C# 正则表达式问题

Question

I am trying to parse the following line:

"\#" TEST #comment hello world

In my input, the #comment always comes at the end of the line. There may or may not be a comment, but if there is, its always in the end of the line.

I used the following Regex to parse it:

(\#.+)?

I have the RegexOption.RightToLeft on. I expected it to pull #comment hello world . But instead it is pulling "#" TEST #comment hello world"

Why is my Regex expression not pulling the right thing and what is the valid Regex expression I need to make it pull correctly?

Answer 1

The important question is: How do you see the difference between the # at the end of the line and the # that starts the comment? Let's assume for simplicity that the last # starts a comment.

In that case, what you want to match is

• one #
• an arbitrary sequence of text not containing #
• until the end of the line

So let's put that into a regex: #[^#]*$ . You don't need RightToLeft for it. As far as I know, you also don't need to escape # in C# regular expressions.

Of course, if you provide information on how to see the difference between a "valid" # and a "comment-starting" #, a more elegant solution could be found that allows for # within comments.

Answer 2

The + operator tries to match as many times as it can. To match as few times as possible, use its lazy equivalent, +? :

(#.+?)

Of course, this would give trouble with comments that contain # :

"\#" TEST #comment #hello #world

Answer 3

Use " #.+". I left the \ out of my test because # is not a recognized escape sequence. I left out the (, ) and? because they where not needed.

Regex regex = new Regex(" #.+");
Console.WriteLine(regex.Match("#\" TEST #comment hello world"));

Answer 4

For the test string you've given, this regex pulls the comment correctly (with right to left option): /((?: #).+)$/

Disclaimer:

• Also pulls the whitespace just before the '#', so you may need to do a trim.
• Comment cannot contain the sequence ' #' in them

Answer 5

This will match "#" and everything after it, witch is the expected behavior:)

var reg = new Regex("#(.)*")

Hope this helps

Answer 6

Right, I've tested this one and it seems to do the necessary.

\#.+(\#.+)$

Specifically, it skips past the first #, then captures everything from the second # to the end of the line, returning

#comment hello world 

Answer 7

I think you'll find too many edge cases when trying to pull this off with regular expressions. Dealing with the quotes is what really complicates things, not to mention escape characters.

A procedural solution is not complicated, and will be faster and easier to modify as needs dictate. Note that I don't know what the escape characters should be in your example, but you could certainly add that to the algorithm...

string CodeSnippet = Resource1.CodeSnippet;
StringBuilder CleanCodeSnippet = new StringBuilder();
bool InsideQuotes = false;
bool InsideComment = false;

Console.WriteLine("BEFORE");
Console.WriteLine(CodeSnippet);
Console.WriteLine("");

for (int i = 0; i < CodeSnippet.Length; i++)
{
    switch(CodeSnippet[i])
    {
        case '"' : 
            if (!InsideComment) InsideQuotes = !InsideQuotes;
            break;
        case '#' :
            if (!InsideQuotes) InsideComment = true;
            break;
        case '\n' :
            InsideComment = false;
            break;                       
    }

    if (!InsideComment)
    {
        CleanCodeSnippet.Append(CodeSnippet[i]);
    }
}

Console.WriteLine("AFTER");
Console.WriteLine(CleanCodeSnippet.ToString());
Console.WriteLine("");

This example strips the comments away from the CodeSnippet . I assumed that's what you were after.

Here's the output:

BEFORE
"\#" TEST #comment hello world
"ab" TEST #comment hello world
"ab" TEST #comment "hello world
"ab" + "ca" + TEST #comment
"\#" TEST
"ab" TEST

AFTER
"\#" TEST
"ab" TEST
"ab" TEST
"ab" + "ca" + TEST
"\#" TEST
"ab" TEST

As I said, you'll probably need to add escape characters to the algorithm. But this is a good starting point.