VB.net的RegEx

Question

I have a txt file with content

$NETS  
P3V3_AUX_LGATE;  PQ6.8 PU37.2   
U335_PIN1;  R3328.1 U335.1  
$END  

need to be updated in this format, and save back to another txt file

$NETS  
'P3V3_AUX_LGATE';  PQ6.8 PU37.2  
'U335_PIN1';  R3328.1 U335.1  
$END

NOTE: number of lines may go up to 10,000 lines

My current solution is to read the txt file line by line, detect the presence of the ";" and newline character and do the changes.

Right now i have a variable that holds ALL the lines, is there other way something like Replace via RegEx to do the changes without looping thru each line, this way i can readily print the result

and follow up question, which one is more efficient?

Answer 1

You could probably find all the matches using something like \\w+; but I don't know how you'd be able to do a replace on that using Regex.Replace to add the ' s but keep the original match.

However, if you already have it as one variable, you don't have to read the file again, either you could make your code find all ; s and then find the previous newline for each, or you could use a String.Split on newlines to split the variable you've already got into lines. And if you want to get it back to one variable you can just use String.Join .

Personally I'd normally use the String.Split (and possibly the String.Join if needed) method, since I think that would make the code easy to read.

Answer 2

Try

ResultString = Regex.Replace(SubjectString, "^([^;\r\n]+);", "'$1';", RegexOptions.Multiline)

on your multiline string.

This will find any string (length one or more) at the start of a line up until the first semicolon if there is one and replace it with its quoted equivalent.

It should be more efficient than looping through the string line by line as you're doing now, but if you're in doubt, you'd have to profile it.

Answer 3

I would say Yes! this can be done with Regular expressions. Make sure you got the "multiline" option turned on and craft your regular expression using some capture groups to ease the work.

I can however say this will NOT be the optimal one. Since you mention the amount of lines you could be processing, it seems 'resource wise' smarter to use a streaming approach instead of the in memory approach.

Taking the Regex approach (and this took 15 mins so please don't think this is an optimal solution, just prove it would work)

    private static Regex matcher = new Regex(@"^\$NETS\r\n(?<entrytitle>.[^;]*);\s*(?<entryrest>.*)\r\n(?<entrytitle2>.[^;]*);\s*(?<entryrest2>.*)\r\n\$END\r\n", RegexOptions.Compiled | RegexOptions.Multiline);
    static void Main(string[] args)
    {
        string newString = matcher.Replace(ExampleFileContent, new MatchEvaluator(evaluator));
    }

    static string evaluator(Match m)
    {
        return String.Format("$NETS\r\n'{0}'; {1}\r\n'{2}'; {3}\r\n$END\r\n",
                              m.Groups["entrytitle"].Value,
                              m.Groups["entryrest"].Value,
                              m.Groups["entrytitle2"].Value,
                              m.Groups["entryrest2"].Value);            
    }

Hope this helps,