使用正则表达式替换模式替换字符串

Question

Basically I am dealing with CSV file and reading it line by line in C#. I have a string input(a line) and trying to find a Regex pattern and replace it using another Regex pattern but result is not what I expect.

var input = "\"efgh ,ijkl123,\",abcd ,  \"efgh ,ijkl123,\",mnop456 \"efgh ,ijkl123,\"";

In output I need to replace internal commas between double quotations with semicolon where those double quotation are between commas themselves.

between double quotation and external comma (comma outside of pair of double quotes) it can be only white space.

So I expect output to be: "efgh ;ijkl123,",abcd , "efgh ;ijkl123,",mnop456 "efgh ,ijkl123,"

my code:

var pattern = @".*,\s*""(.*,+.*)+""\s*,.*";
var replacePattern = @".*,\s*""(.*;+.*)+""\s*,.*";
if (Regex.IsMatch(input, pattern))
{
    var output = Regex.Replace(input, pattern, replacePattern);
}

but running my code, output is: . ,\\s "(. ;+. )+"\\s*,.* which is replacePattern.

EDIT more input sample and output as expected:

1. input abcd , "efgh ,ijkl123,",mnop456

   output abcd , "efgh ;ijkl123;",mnop456

2. input "efgh ,ijkl123,",abcd , "efgh ,ijkl123,",mnop456 "efgh ,ijkl123,"

   output "efgh ;ijkl123;",abcd , "efgh ;ijkl123;",mnop456 "efgh ,ijkl123,"

3. input ,"efgh ,ijkl123,",abcd" , "efgh ijkl123,",mnop456 "efgh ,ijkl123,","efgh ,ijkl123,"mnop456

   output ,"efgh ;ijkl123;",abcd" , "efgh ijkl123;",mnop456 "efgh ,ijkl123,","efgh ,ijkl123,"mnop456

4. input ,"efgh" ,ijkl123,",abcd" , "efgh ijkl123,",mnop456 "efgh ,ijkl123,","efgh ,ijkl123,"mnop456

   output ,"efgh" ,ijkl123,";abcd" , "efgh ijkl123;",mnop456 "efgh ,ijkl123,","efgh ,ijkl123,"mnop456

5. input efgh ,ijkl123,",abcd , "efgh ,ijkl123,",mnop456 "efgh ,ijkl123,"

   output efgh ,ijkl123,",abcd , "efgh ;ijkl123;",mnop456 "efgh ,ijkl123,"

Answer 1

Well, this is bit tricky and I'm sure someone will suggest a better regex than mine. Suppose you input text is:

"efgh ,ijkl123,",abcd ,  "efgh ,ijkl123,",mnop456 "efgh ,ijkl123,"

You can try:

var data = "\"efgh ,ijkl123,\",abcd ,  \"efgh ,ijkl123,\",mnop456 \"efgh ,ijkl123,\"";

var rx = @"(?<=(^|,[ \t]*))\""[^\""\n]+\""(?=[ \t]*(,|$))";

var matches = Regex.Matches (data, rx);

foreach (Match match in matches) {
    data = new Regex (match.Value).
        Replace(data, match.Value.Replace (',', ';'), 1);
}

Console.WriteLine (data);

It will emit:

"efgh ;ijkl123;",abcd ,  "efgh ;ijkl123;",mnop456, "efgh ,ijkl123," 

Code above is essentially replacing all , comas between double quotes with ; semi colons.

Answer 2

Not sure is it very efficient, but works. Suggestions are welcome to improve it further.

string  input = "\"efgh ,ijkl123,\",abcd ,  \"efgh ,ijkl123,\",mnop456 \"efgh ,ijkl123,\"";; 

Regex.Matches(input, "\"([^\"]*)\"(,)") // Extract string between quotes followed by ','.
.Cast<Match>()
    .ToList()
    .ForEach(m=> input = input.Replace(m.Value, m.Value.Replace(",",";")) // for each match replace with ';' inserted match.
                              .Replace(";\";",",\","));  // a hack, should have done it better

Ouput :

"efgh ;ijkl123,",abcd ,  "efgh ;ijkl123,",mnop456 "efgh ,ijkl123,"

Working Demo