Develop Reference

Split string at semicolon get second part only

I have a string in the following format:
message:action
How can I parse this string and get the second part only?
The action might be different the first part will always be the same: message:
Is this a correct way?
string value="message:action";
string[] result = value.Split(':');
string action = result[1];

Yes it is. Or, since the first part is always the same, you could simply use substring:
value.SubString(8);

That solution is fine, but it's a little more work than necessary. If you know that the string will always start with "message:", why not just do this:
var action = value.Substring(8);
However, if the string that comes before the : might change, but you still only care about what comes after it, you could do this:
var action = value.Substring(value.IndexOf(':') + 1);

Yes, your solution will work. However, if the first part is always the same, another possibility is to use Substring as pswg's answer mentions.

If you know that the string will start with message: you can do this:
String value="message:action";
String result = value.Replace("message:","");

It's OK, although if you worry about there being colons in the content (e.g., "message:action:do this"):
string value="message:action";
string[] result = value.Split(':');
string action = value.Substring(result[0].Length);
(this is of course if it's not always starting with "message:")

If you want to have self-documenting code, you can use regex (so you don't think in the future, what the heck is this code and why do I take second part).
string value = "message:action";
string action = Regex.Match(value, "(?<=^message:).*").Value; // will throw exception in case if someday format changes

I'm adding one more answer for legibility. There are several correct answers in the thread so far.
Just to offer a very readable alternative at very little cost of performance consider the following:
string value = "message:action";
string action = value.Split(':').Skip(1);

Correction in this simple regular expression

I am new to regular expressions and the one that i have written might be a very simple one but donot know where I am wrong.
#"^([a-zA-Z._]+)#([\d]+)"
This RE is for the following string:
somename#somenumber
Now i am trying to retrieve the somename and somenumber. This is what i did:
ac.name = m.Groups[0].Value;
ac.number = m.Groups[1].Value;
Here ac.name reads the complete string, and ac.number reads somenumber. Where am I wrong in ac.name?

i guess the regex is correct, the problem is, you get the ac.name not from group 1 but group(0), which is the whole string. try this:
ac.name = m.Groups[1].Value;
ac.number = m.Groups[2].Value;

This regex is correct. I think your mistake is in somewhere else. You seem to use C#. So, you should think about the regex usage in the language.
Looking to the code sample in MSDN, you need to use 1-based indexes while accessing Groups instead of zero-based (as also Kent suggested). So, use this:
String name = m.Groups[1].Value;
String number = m.Groups[2].Value;

use this regex (\w+)#(\d+([.,]\d+)?)
Groups[1] will be contain name
Groups[2] will be contain number

I think you should move the + into the capture group:
#"^([a-zA-Z._]+)#([\d]+)"

If this is C#, try without the ^
([a-zA-Z\._]+)#([\d]+)
I just tried it out and it groups properly
Update: escaped the .
If you want only one match (and hence the ^ in original expression), use .Match instead of .Matches method. See MSDN documentation on Regular Expression Classes.

Trimming strings between two points

I have a string as follows:
ListViewSubItem: {Debian6/Debian6.vmx }
What would be the most efficient way of getting the following output:
Debian6/Debian6.vmx

int beginidx = haystack.IndexOf('{');
string needle = haystack.SubString(beginidx,
haystack.IndexOf('}') - beginidx + 1).Trim();

string result = Regex.Match("ListViewSubItem: {Debian6/Debian6.vmx }", #"(?<={)(.+?)(?=})").Value;

You can use a regex:
\{\s*(.*)\s*\}
The desired string will be in the first captured group (Match.Groups[1]).
Example:
string output = Regex.Match(input, #"\{\s*(.*?)\s*\}").Groups[1].Value;
As was pointed out, regexes are slower than plain string manipulation. If performance is an issue, and the string extraction is in a tight loop, then it may be better to use an optimized method. Otherwise, regex vs string is IMHO a matter of personal preference.

I'm going to guess you didn't generate that string yourself. So the question is, why parse it? This looks like the sort of question where you should take a step back and think of what you're trying to solve, and if another method might be cleaner.
I am not familiar with the ListViewSubItem class, but it seems like you have one of these and you have called ToString() on it. Now you are parsing the output of ToString() to see the model object your sub-item represents.
Does this not seem backwards to you? Maybe you should deal with the ListViewSubItem object itself (from a brief look at the documentation, what you want seems to be the Text property), rather than fiddling with what ToString() returns, which seems volatile and dependent on the implementation of that class's ToString() method (which, though likely not, could theoretically change from release to release). Not to mention corner cases like "what if the string contains the } character?"
For this in other reasons, as a general rule I think you should not have to deal with serialization when your data source is in memory the whole time.
So, in summary, if you have something like this:
ListViewSubItem item = /* ... */;
string s = item.ToString();
// TODO: parse {Debian} out of ListViewSubItem {Debian}
Why not this instead:
ListViewSubItem item = /* ... */;
string OS = item.Text;

about string removing in C#

Whats is correct to do?
check if exists, then remove?
var input = "foo #main baa";
if(input.Contains("#main")) {
input = input.Replace("#main", "");
}
or just:
input = input.Replace("#main", "");
Well, this seem a simple question,but I really want know.
Thanks in advance.

The Contains check actually just makes your code slower.
Remove it.
The Contains call needs to loop through the string until it finds #main.
The Replace call then needs to do the same exact loop (it can't remember it from the Contains call).
This is a Shlemiel the Painter's algorithm.
Replace can handle strings with zero or more occurrences of the search string, so you don't need the check.

Just do the replacement - if it's not there, nothing should happen.

Just make the call to Replace(). If the substring isn't found nothing happens and you avoid an additional call to Contains().

I would do this:
input = input.Replace("#main", "").Replace(" "," ");
To remove any double spaces.

Just remove it. The only thing to check is if the string is null or not.

C# Extracting a name from a string

I want to extract 'James\, Brown' from the string below but I don't always know what the name will be. The comma is causing me some difficuly so what would you suggest to extract James\, Brown?
OU=James\, Brown,OU=Test,DC=Internal,DC=Net
Thanks

A regex is likely your best approach
static string ParseName(string arg) {
var regex = new Regex(#"^OU=([a-zA-Z\\]+\,\s+[a-zA-Z\\]+)\,.*$");
var match = regex.Match(arg);
return match.Groups[1].Value;
}

You can use a regex:
string input = #"OU=James\, Brown,OU=Test,DC=Internal,DC=Net";
Match m = Regex.Match(input, "^OU=(.*?),OU=.*$");
Console.WriteLine(m.Groups[1].Value);

A quite brittle way to do this might be...
string name = #"OU=James\, Brown,OU=Test,DC=Internal,DC=Net";
string[] splitUp = name.Split("=".ToCharArray(),3);
string namePart = splitUp[1].Replace(",OU","");
Console.WriteLine(namePart);
I wouldn't necessarily advocate this method, but I've just come back from a departmental Christmas lyunch and my brain is not fully engaged yet.

I'd start off with a regex to split up the groups:
Regex rx = new Regex(#"(?<!\\),");
String test = "OU=James\\, Brown,OU=Test,DC=Internal,DC=Net";
String[] segments = rx.Split(test);
But from there I would split up the parameters in the array by splitting them up manually, so that you don't have to use a regex that depends on more than the separator character used. Since this looks like an LDAP query, it might not matter if you always look at params[0], but there is a chance that the name might be set as "CN=". You can cover both cases by just reading the query like this:
String name = segments[0].Split('=', 2)[1];

That looks suspiciously like an LDAP or Active Directory distinguished name formatted according to RFC 2253/4514.
Unless you're working with well known names and/or are okay with a fragile hackaround (like the regex solutions) - then you should start by reading the spec.
If you, like me, generally hate implementing code according to RFCs - then hope this guy did a better job following the spec than you would. At least he claims to be 2253 compliant.

If the slash is always there, I would look at potentially using RegEx to do the match, you can use a match group for the last and first names.
^OU=([a-zA-Z])\,\s([a-zA-Z])
That RegEx will match names that include characters only, you will need to refine it a bit for better matching for the non-standard names. Here is a RegEx tester to help you along the way if you go this route.

Replace \, with your own preferred magic string (perhaps & #44;), split on remaining commas or search til the first comma, then replace your magic string with a single comma.
i.e. Something like:
string originalStr = #"OU=James\, Brown,OU=Test,DC=Internal,DC=Net";
string replacedStr = originalStr.Replace("\,", ",");
string name = replacedStr.Substring(0, replacedStr.IndexOf(","));
Console.WriteLine(name.Replace(",", ","));

Assuming you're running in Windows, use PInvoke with DsUnquoteRdnValueW. For code, see my answer to another question: https://stackoverflow.com/a/11091804/628981

If the format is always the same:
string line = GetStringFromWherever();
int start = line.IndexOf("=") + 1;//+1 to get start of name
int end = line.IndexOf("OU=",start) -1; //-1 to remove comma
string name = line.Substring(start, end - start);
Forgive if syntax is not quite right - from memory. Obviously this is not very robust and fails if the format ever changes.