Regex not matching in C#

Question

I have the following string:

 string error = "<MESSAGES><MESSAGE SEVERITY=\"2\" NUMBER=\"16\" TEXT=\"The Record Case is locked by user\" /></MESSAGES>";

I want to match between the TEXT=\" and the following \

I'm using the following expression var regex = new Regex(@"TEXT=\\""(.*?)\\");

Expresso tells me this regex is correct. RegExr tells me this regex is correct.

But C# disagrees.

I've tried

• Groups[] and match.Value.
• \x22 instead of " as I thought it might be an escape problem.
• /TEXT=\""(.*?)\/g

All, to no avail.

What am I missing?

Answer 1

Use XElement, you have an XML fragment:

var error = "<MESSAGES><MESSAGE SEVERITY=\"2\" NUMBER=\"16\" TEXT=\"The Record Case is locked by user\" /></MESSAGES>";
var xe = XElement.Parse(error);
var res = xe.Elements("MESSAGE")
                   .Where(p => p.HasAttributes && p.Attributes("TEXT") != null)
                   .Select(n => n.Attribute("TEXT").Value)
                   .ToList();

Output:

Mind that with very large input strings, .*? may cause catastrophic backtracking, that is why you should avoid using it whenever possible. If you need a regex for this (because some of your input is not XML-valid), you can use:

var attr_vals = Regex.Matches(error, @"(?i)\bTEXT=""([^""]+)""")
             .OfType<Match>()
             .Select(p => p.Groups[1].Value)
             .ToList();

(2 times faster than Karthik's, tested on regexhero.com)

Output:

Mind that with regex, you will get all XML entities untouched (e.g. & and not &). You will have to use System.Web.HttpUtility later.

Answer 2

Use the following (your actual string will be compiled to a string without \'s.. since you are just using them as escape characters):

var regex = new Regex(@"TEXT=""([^""]+)""");

Answer 3

This works for me:

Regex.Match(error, "TEXT=\\\"(.*?)\\\"")

You need to escape both \ and " character with \