Edit

Share via


Quantifiers in Regular Expressions

Quantifiers specify how many instances of a character, group, or character class must be present in the input for a match to be found. The following table lists the quantifiers supported by .NET:

Greedy quantifier Lazy quantifier Description
* *? Matches zero or more times.
+ +? Matches one or more times.
? ?? Matches zero or one time.
{ n } { n }? Matches exactly n times.
{ n ,} { n ,}? Matches at least n times.
{ n , m } { n , m }? Matches from n to m times.

The quantities n and m are integer constants. Ordinarily, quantifiers are greedy. They cause the regular expression engine to match as many occurrences of particular patterns as possible. Appending the ? character to a quantifier makes it lazy. It causes the regular expression engine to match as few occurrences as possible. For a complete description of the difference between greedy and lazy quantifiers, see the section Greedy and Lazy Quantifiers later in this article.

Important

Nesting quantifiers, such as the regular expression pattern (a*)*, can increase the number of comparisons that the regular expression engine must perform. The number of comparisons can increase as an exponential function of the number of characters in the input string. For more information about this behavior and its workarounds, see Backtracking.

Regular Expression Quantifiers

The following sections list the quantifiers supported by .NET regular expressions:

Note

If the *, +, ?, {, and } characters are encountered in a regular expression pattern, the regular expression engine interprets them as quantifiers or part of quantifier constructs unless they are included in a character class. To interpret these as literal characters outside a character class, you must escape them by preceding them with a backslash. For example, the string \* in a regular expression pattern is interpreted as a literal asterisk ("*") character.

Match Zero or More Times: *

The * quantifier matches the preceding element zero or more times. It's equivalent to the {0,} quantifier. * is a greedy quantifier whose lazy equivalent is *?.

The following example illustrates this regular expression. Five of the nine digit-groups in the input string match the pattern and four (95, 929, 9219, and 9919) don't.

string pattern = @"\b91*9*\b";
string input = "99 95 919 929 9119 9219 999 9919 91119";
foreach (Match match in Regex.Matches(input, pattern))
   Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index);

// The example displays the following output:
//       '99' found at position 0.
//       '919' found at position 6.
//       '9119' found at position 14.
//       '999' found at position 24.
//       '91119' found at position 33.
Dim pattern As String = "\b91*9*\b"
Dim input As String = "99 95 919 929 9119 9219 999 9919 91119"
For Each match As Match In Regex.Matches(input, pattern)
    Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index)
Next
' The example displays the following output:
'       '99' found at position 0.
'       '919' found at position 6.
'       '9119' found at position 14.
'       '999' found at position 24.
'       '91119' found at position 33.

The regular expression pattern is defined as shown in the following table:

Pattern Description
\b Specifies that the match must start at a word boundary.
91* Matches a 9 followed by zero or more 1 characters.
9* Matches zero or more 9 characters.
\b Specifies that the match must end at a word boundary.

Match One or More Times: +

The + quantifier matches the preceding element one or more times. It's equivalent to {1,}. + is a greedy quantifier whose lazy equivalent is +?.

For example, the regular expression \ban+\w*?\b tries to match entire words that begin with the letter a followed by one or more instances of the letter n. The following example illustrates this regular expression. The regular expression matches the words an, annual, announcement, and antique, and correctly fails to match autumn and all.

string pattern = @"\ban+\w*?\b";

string input = "Autumn is a great time for an annual announcement to all antique collectors.";
foreach (Match match in Regex.Matches(input, pattern, RegexOptions.IgnoreCase))
   Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index);

// The example displays the following output:
//       'an' found at position 27.
//       'annual' found at position 30.
//       'announcement' found at position 37.
//       'antique' found at position 57.
Dim pattern As String = "\ban+\w*?\b"

Dim input As String = "Autumn is a great time for an annual announcement to all antique collectors."
For Each match As Match In Regex.Matches(input, pattern, RegexOptions.IgnoreCase)
    Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index)
Next
' The example displays the following output:
'       'an' found at position 27.
'       'annual' found at position 30.
'       'announcement' found at position 37.
'       'antique' found at position 57.

The regular expression pattern is defined as shown in the following table:

Pattern Description
\b Start at a word boundary.
an+ Matches an a followed by one or more n characters.
\w*? Matches a word character zero or more times but as few times as possible.
\b End at a word boundary.

Match Zero or One Time: ?

The ? quantifier matches the preceding element zero or one time. It's equivalent to {0,1}. ? is a greedy quantifier whose lazy equivalent is ??.

For example, the regular expression \ban?\b tries to match entire words that begin with the letter a followed by zero or one instance of the letter n. In other words, it tries to match the words a and an. The following example illustrates this regular expression:

string pattern = @"\ban?\b";
string input = "An amiable animal with a large snout and an animated nose.";
foreach (Match match in Regex.Matches(input, pattern, RegexOptions.IgnoreCase))
   Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index);

// The example displays the following output:
//        'An' found at position 0.
//        'a' found at position 23.
//        'an' found at position 42.
Dim pattern As String = "\ban?\b"
Dim input As String = "An amiable animal with a large snout and an animated nose."
For Each match As Match In Regex.Matches(input, pattern, RegexOptions.IgnoreCase)
    Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index)
Next
' The example displays the following output:
'       'An' found at position 0.
'       'a' found at position 23.
'       'an' found at position 42.

The regular expression pattern is defined as shown in the following table:

Pattern Description
\b Start at a word boundary.
an? Matches an a followed by zero or one n character.
\b End at a word boundary.

Match Exactly n Times: {n}

The {n} quantifier matches the preceding element exactly n times, where n is any integer. {n} is a greedy quantifier whose lazy equivalent is {n}?.

For example, the regular expression \b\d+\,\d{3}\b tries to match a word boundary followed by one or more decimal digits followed by three decimal digits followed by a word boundary. The following example illustrates this regular expression:

string pattern = @"\b\d+\,\d{3}\b";
string input = "Sales totaled 103,524 million in January, " +
                      "106,971 million in February, but only " +
                      "943 million in March.";
foreach (Match match in Regex.Matches(input, pattern))
   Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index);

//  The example displays the following output:
//        '103,524' found at position 14.
//        '106,971' found at position 45.
Dim pattern As String = "\b\d+\,\d{3}\b"
Dim input As String = "Sales totaled 103,524 million in January, " + _
                      "106,971 million in February, but only " + _
                      "943 million in March."
For Each match As Match In Regex.Matches(input, pattern)
    Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index)
Next
' The example displays the following output:
'       '103,524' found at position 14.
'       '106,971' found at position 45.

The regular expression pattern is defined as shown in the following table:

Pattern Description
\b Start at a word boundary.
\d+ Matches one or more decimal digits.
\, Matches a comma character.
\d{3} Matches three decimal digits.
\b End at a word boundary.

Match at Least n Times: {n,}

The {n,} quantifier matches the preceding element at least n times, where n is any integer. {n,} is a greedy quantifier whose lazy equivalent is {n,}?.

For example, the regular expression \b\d{2,}\b\D+ tries to match a word boundary followed by at least two digits followed by a word boundary and a non-digit character. The following example illustrates this regular expression. The regular expression fails to match the phrase "7 days" because it contains just one decimal digit, but it successfully matches the phrases "10 weeks" and "300 years".

string pattern = @"\b\d{2,}\b\D+";
string input = "7 days, 10 weeks, 300 years";
foreach (Match match in Regex.Matches(input, pattern))
   Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index);

//  The example displays the following output:
//        '10 weeks, ' found at position 8.
//        '300 years' found at position 18.
Dim pattern As String = "\b\d{2,}\b\D+"
Dim input As String = "7 days, 10 weeks, 300 years"
For Each match As Match In Regex.Matches(input, pattern)
    Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index)
Next
' The example displays the following output:
'       '10 weeks, ' found at position 8.
'       '300 years' found at position 18.

The regular expression pattern is defined as shown in the following table:

Pattern Description
\b Start at a word boundary.
\d{2,} Matches at least two decimal digits.
\b Matches a word boundary.
\D+ Matches at least one non-decimal digit.

Match Between n and m Times: {n,m}

The {n,m} quantifier matches the preceding element at least n times, but no more than m times, where n and m are integers. {n,m} is a greedy quantifier whose lazy equivalent is {n,m}?.

In the following example, the regular expression (00\s){2,4} tries to match between two and four occurrences of two zero digits followed by a space. The final portion of the input string includes this pattern five times rather than the maximum of four. However, only the initial portion of this substring (up to the space and the fifth pair of zeros) matches the regular expression pattern.

string pattern = @"(00\s){2,4}";
string input = "0x00 FF 00 00 18 17 FF 00 00 00 21 00 00 00 00 00";
foreach (Match match in Regex.Matches(input, pattern))
   Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index);

//  The example displays the following output:
//        '00 00 ' found at position 8.
//        '00 00 00 ' found at position 23.
//        '00 00 00 00 ' found at position 35.
Dim pattern As String = "(00\s){2,4}"
Dim input As String = "0x00 FF 00 00 18 17 FF 00 00 00 21 00 00 00 00 00"
For Each match As Match In Regex.Matches(input, pattern)
    Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index)
Next
' The example displays the following output:
'       '00 00 ' found at position 8.
'       '00 00 00 ' found at position 23.
'       '00 00 00 00 ' found at position 35.

Match Zero or More Times (Lazy Match): *?

The *? quantifier matches the preceding element zero or more times but as few times as possible. It's the lazy counterpart of the greedy quantifier *.

In the following example, the regular expression \b\w*?oo\w*?\b matches all words that contain the string oo.

 string pattern = @"\b\w*?oo\w*?\b";
 string input = "woof root root rob oof woo woe";
 foreach (Match match in Regex.Matches(input, pattern, RegexOptions.IgnoreCase))
    Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index);

 //  The example displays the following output:
//        'woof' found at position 0.
//        'root' found at position 5.
//        'root' found at position 10.
//        'oof' found at position 19.
//        'woo' found at position 23.
Dim pattern As String = "\b\w*?oo\w*?\b"
Dim input As String = "woof root root rob oof woo woe"
For Each match As Match In Regex.Matches(input, pattern, RegexOptions.IgnoreCase)
    Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index)
Next
' The example displays the following output:
'       'woof' found at position 0.
'       'root' found at position 5.
'       'root' found at position 10.
'       'oof' found at position 19.
'       'woo' found at position 23.

The regular expression pattern is defined as shown in the following table:

Pattern Description
\b Start at a word boundary.
\w*? Matches zero or more word characters but as few characters as possible.
oo Matches the string oo.
\w*? Matches zero or more word characters but as few characters as possible.
\b End on a word boundary.

Match One or More Times (Lazy Match): +?

The +? quantifier matches the preceding element one or more times but as few times as possible. It's the lazy counterpart of the greedy quantifier +.

For example, the regular expression \b\w+?\b matches one or more characters separated by word boundaries. The following example illustrates this regular expression:

string pattern = @"\b\w+?\b";
string input = "Aa Bb Cc Dd Ee Ff";
foreach (Match match in Regex.Matches(input, pattern))
   Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index);

//  The example displays the following output:
//        'Aa' found at position 0.
//        'Bb' found at position 3.
//        'Cc' found at position 6.
//        'Dd' found at position 9.
//        'Ee' found at position 12.
//        'Ff' found at position 15.
Dim pattern As String = "\b\w+?\b"
Dim input As String = "Aa Bb Cc Dd Ee Ff"
For Each match As Match In Regex.Matches(input, pattern)
    Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index)
Next
' The example displays the following output:
'       'Aa' found at position 0.
'       'Bb' found at position 3.
'       'Cc' found at position 6.
'       'Dd' found at position 9.
'       'Ee' found at position 12.
'       'Ff' found at position 15.

Match Zero or One Time (Lazy Match): ??

The ?? quantifier matches the preceding element zero or one time but as few times as possible. It's the lazy counterpart of the greedy quantifier ?.

For example, the regular expression ^\s*(System.)??Console.Write(Line)??\(?? attempts to match the strings Console.Write or Console.WriteLine. The string can also include System. before Console, and it can be followed by an opening parenthesis. The string must be at the beginning of a line, although it can be preceded by white space. The following example illustrates this regular expression:

string pattern = @"^\s*(System.)??Console.Write(Line)??\(??";
string input = "System.Console.WriteLine(\"Hello!\")\n" +
                      "Console.Write(\"Hello!\")\n" +
                      "Console.WriteLine(\"Hello!\")\n" +
                      "Console.ReadLine()\n" +
                      "   Console.WriteLine";
foreach (Match match in Regex.Matches(input, pattern,
                                      RegexOptions.IgnorePatternWhitespace |
                                      RegexOptions.IgnoreCase |
                                      RegexOptions.Multiline))
   Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index);

//  The example displays the following output:
//        'System.Console.Write' found at position 0.
//        'Console.Write' found at position 36.
//        'Console.Write' found at position 61.
//        '   Console.Write' found at position 110.
Dim pattern As String = "^\s*(System.)??Console.Write(Line)??\(??"
Dim input As String = "System.Console.WriteLine(""Hello!"")" + vbCrLf + _
                      "Console.Write(""Hello!"")" + vbCrLf + _
                      "Console.WriteLine(""Hello!"")" + vbCrLf + _
                      "Console.ReadLine()" + vbCrLf + _
                      "   Console.WriteLine"
For Each match As Match In Regex.Matches(input, pattern, _
                                         RegexOptions.IgnorePatternWhitespace Or RegexOptions.IgnoreCase Or RegexOptions.MultiLine)
    Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index)
Next
' The example displays the following output:
'       'System.Console.Write' found at position 0.
'       'Console.Write' found at position 36.
'       'Console.Write' found at position 61.
'       '   Console.Write' found at position 110.

The regular expression pattern is defined as shown in the following table:

Pattern Description
^ Matches the start of the input stream.
\s* Matches zero or more white-space characters.
(System.)?? Matches zero or one occurrence of the string System..
Console.Write Matches the string Console.Write.
(Line)?? Matches zero or one occurrence of the string Line.
\(?? Matches zero or one occurrence of the opening parenthesis.

Match Exactly n Times (Lazy Match): {n}?

The {n}? quantifier matches the preceding element exactly n times, where n is any integer. It's the lazy counterpart of the greedy quantifier {n}.

In the following example, the regular expression \b(\w{3,}?\.){2}?\w{3,}?\b is used to identify a website address. The expression matches www.microsoft.com and msdn.microsoft.com but doesn't match mywebsite or mycompany.com.

string pattern = @"\b(\w{3,}?\.){2}?\w{3,}?\b";
string input = "www.microsoft.com msdn.microsoft.com mywebsite mycompany.com";
foreach (Match match in Regex.Matches(input, pattern))
   Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index);

//  The example displays the following output:
//        'www.microsoft.com' found at position 0.
//        'msdn.microsoft.com' found at position 18.
Dim pattern As String = "\b(\w{3,}?\.){2}?\w{3,}?\b"
Dim input As String = "www.microsoft.com msdn.microsoft.com mywebsite mycompany.com"
For Each match As Match In Regex.Matches(input, pattern)
    Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index)
Next
' The example displays the following output:
'       'www.microsoft.com' found at position 0.
'       'msdn.microsoft.com' found at position 18.

The regular expression pattern is defined as shown in the following table:

Pattern Description
\b Start at a word boundary.
(\w{3,}?\.) Matches at least three word-characters but as few characters as possible, followed by a dot or period character. This pattern is the first capturing group.
(\w{3,}?\.){2}? Matches the pattern in the first group two times but as few times as possible.
\b End the match on a word boundary.

Match at Least n Times (Lazy Match): {n,}?

The {n,}? quantifier matches the preceding element at least n times, where n is any integer but as few times as possible. It's the lazy counterpart of the greedy quantifier {n,}.

See the example for the {n}? quantifier in the previous section for an illustration. The regular expression in that example uses the {n,} quantifier to match a string that has at least three characters followed by a period.

Match Between n and m Times (Lazy Match): {n,m}?

The {n,m}? quantifier matches the preceding element between n and m times, where n and m are integers but as few times as possible. It's the lazy counterpart of the greedy quantifier {n,m}.

In the following example, the regular expression \b[A-Z](\w*?\s*?){1,10}[.!?] matches sentences that contain between 1 and 10 words. It matches all the sentences in the input string except for one sentence that contains 18 words.

string pattern = @"\b[A-Z](\w*?\s*?){1,10}[.!?]";
string input = "Hi. I am writing a short note. Its purpose is " +
                      "to test a regular expression that attempts to find " +
                      "sentences with ten or fewer words. Most sentences " +
                      "in this note are short.";
foreach (Match match in Regex.Matches(input, pattern))
   Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index);

//  The example displays the following output:
//        'Hi.' found at position 0.
//        'I am writing a short note.' found at position 4.
//        'Most sentences in this note are short.' found at position 132.
Dim pattern As String = "\b[A-Z](\w*\s?){1,10}?[.!?]"
Dim input As String = "Hi. I am writing a short note. Its purpose is " + _
                      "to test a regular expression that attempts to find " + _
                      "sentences with ten or fewer words. Most sentences " + _
                      "in this note are short."
For Each match As Match In Regex.Matches(input, pattern)
    Console.WriteLine("'{0}' found at position {1}.", match.Value, match.Index)
Next
' The example displays the following output:
'       'Hi.' found at position 0.
'       'I am writing a short note.' found at position 4.
'       'Most sentences in this note are short.' found at position 132.

The regular expression pattern is defined as shown in the following table:

Pattern Description
\b Start at a word boundary.
[A-Z] Matches an uppercase character from A to Z.
(\w*?\s*?) Matches zero or more word characters, followed by one or more white-space characters but as few times as possible. This pattern is the first capturing group.
{1,10} Matches the previous pattern between 1 and 10 times.
[.!?] Matches any one of the punctuation characters ., !, or ?.

Greedy and Lazy Quantifiers

Some quantifiers have two versions:

  • A greedy version.

    A greedy quantifier tries to match an element as many times as possible.

  • A non-greedy (or lazy) version.

    A non-greedy quantifier tries to match an element as few times as possible. You can turn a greedy quantifier into a lazy quantifier by adding a ?.

Consider a regular expression that's intended to extract the last four digits from a string of numbers, such as a credit card number. The version of the regular expression that uses the * greedy quantifier is \b.*([0-9]{4})\b. However, if a string contains two numbers, this regular expression matches the last four digits of the second number only, as the following example shows:

string greedyPattern = @"\b.*([0-9]{4})\b";
string input1 = "1112223333 3992991999";
foreach (Match match in Regex.Matches(input1, greedyPattern))
   Console.WriteLine("Account ending in ******{0}.", match.Groups[1].Value);

// The example displays the following output:
//       Account ending in ******1999.
Dim greedyPattern As String = "\b.*([0-9]{4})\b"
Dim input1 As String = "1112223333 3992991999"
For Each match As Match In Regex.Matches(input1, greedypattern)
    Console.WriteLine("Account ending in ******{0}.", match.Groups(1).Value)
Next
' The example displays the following output:
'       Account ending in ******1999.

The regular expression fails to match the first number because the * quantifier tries to match the previous element as many times as possible in the entire string, and so it finds its match at the end of the string.

This behavior isn't the desired one. Instead, you can use the *? lazy quantifier to extract digits from both numbers, as the following example shows:

string lazyPattern = @"\b.*?([0-9]{4})\b";
string input2 = "1112223333 3992991999";
foreach (Match match in Regex.Matches(input2, lazyPattern))
   Console.WriteLine("Account ending in ******{0}.", match.Groups[1].Value);

// The example displays the following output:
//       Account ending in ******3333.
//       Account ending in ******1999.
Dim lazyPattern As String = "\b.*?([0-9]{4})\b"
Dim input2 As String = "1112223333 3992991999"
For Each match As Match In Regex.Matches(input2, lazypattern)
    Console.WriteLine("Account ending in ******{0}.", match.Groups(1).Value)
Next
' The example displays the following output:
'       Account ending in ******3333.
'       Account ending in ******1999.

In most cases, regular expressions with greedy and lazy quantifiers return the same matches. They most commonly return different results when they're used with the wildcard (.) metacharacter, which matches any character.

Quantifiers and Empty Matches

The quantifiers *, +, and {n,m} and their lazy counterparts never repeat after an empty match when the minimum number of captures has been found. This rule prevents quantifiers from entering infinite loops on empty subexpression matches when the maximum number of possible group captures is infinite or near infinite.

For example, the following code shows the result of a call to the Regex.Match method with the regular expression pattern (a?)*, which matches zero or one a character zero or more times. The single capturing group captures each a and String.Empty, but there's no second empty match because the first empty match causes the quantifier to stop repeating.

using System;
using System.Text.RegularExpressions;

public class Example
{
   public static void Main()
   {
      string pattern = "(a?)*";
      string input = "aaabbb";
      Match match = Regex.Match(input, pattern);
      Console.WriteLine("Match: '{0}' at index {1}",
                        match.Value, match.Index);
      if (match.Groups.Count > 1) {
         GroupCollection groups = match.Groups;
         for (int grpCtr = 1; grpCtr <= groups.Count - 1; grpCtr++) {
            Console.WriteLine("   Group {0}: '{1}' at index {2}",
                              grpCtr,
                              groups[grpCtr].Value,
                              groups[grpCtr].Index);
            int captureCtr = 0;
            foreach (Capture capture in groups[grpCtr].Captures) {
               captureCtr++;
               Console.WriteLine("      Capture {0}: '{1}' at index {2}",
                                 captureCtr, capture.Value, capture.Index);
            }
         }
      }
   }
}
// The example displays the following output:
//       Match: 'aaa' at index 0
//          Group 1: '' at index 3
//             Capture 1: 'a' at index 0
//             Capture 2: 'a' at index 1
//             Capture 3: 'a' at index 2
//             Capture 4: '' at index 3
Imports System.Text.RegularExpressions

Module Example
    Public Sub Main()
        Dim pattern As String = "(a?)*"
        Dim input As String = "aaabbb"
        Dim match As Match = Regex.Match(input, pattern)
        Console.WriteLine("Match: '{0}' at index {1}",
                          match.Value, match.Index)
        If match.Groups.Count > 1 Then
            Dim groups As GroupCollection = match.Groups
            For grpCtr As Integer = 1 To groups.Count - 1
                Console.WriteLine("   Group {0}: '{1}' at index {2}",
                                  grpCtr,
                                  groups(grpCtr).Value,
                                  groups(grpCtr).Index)
                Dim captureCtr As Integer = 0
                For Each capture As Capture In groups(grpCtr).Captures
                    captureCtr += 1
                    Console.WriteLine("      Capture {0}: '{1}' at index {2}",
                                      captureCtr, capture.Value, capture.Index)
                Next
            Next
        End If
    End Sub
End Module
' The example displays the following output:
'       Match: 'aaa' at index 0
'          Group 1: '' at index 3
'             Capture 1: 'a' at index 0
'             Capture 2: 'a' at index 1
'             Capture 3: 'a' at index 2
'             Capture 4: '' at index 3

To see the practical difference between a capturing group that defines a minimum and a maximum number of captures and one that defines a fixed number of captures, consider the regular expression patterns (a\1|(?(1)\1)){0,2} and (a\1|(?(1)\1)){2}. Both regular expressions consist of a single capturing group, which is defined in the following table:

Pattern Description
(a\1 Either matches a along with the value of the first captured group …
|(?(1) … or tests whether the first captured group has been defined. The (?(1) construct doesn't define a capturing group.
\1)) If the first captured group exists, match its value. If the group doesn't exist, the group will match String.Empty.

The first regular expression tries to match this pattern between zero and two times; the second, exactly two times. Because the first pattern reaches its minimum number of captures with its first capture of String.Empty, it never repeats to try to match a\1. The {0,2} quantifier allows only empty matches in the last iteration. In contrast, the second regular expression does match a because it evaluates a\1 a second time. The minimum number of iterations, 2, forces the engine to repeat after an empty match.

using System;
using System.Text.RegularExpressions;

public class Example
{
   public static void Main()
   {
      string pattern, input;

      pattern = @"(a\1|(?(1)\1)){0,2}";
      input = "aaabbb";

      Console.WriteLine("Regex pattern: {0}", pattern);
      Match match = Regex.Match(input, pattern);
      Console.WriteLine("Match: '{0}' at position {1}.",
                        match.Value, match.Index);
      if (match.Groups.Count > 1) {
         for (int groupCtr = 1; groupCtr <= match.Groups.Count - 1; groupCtr++)
         {
            Group group = match.Groups[groupCtr];
            Console.WriteLine("   Group: {0}: '{1}' at position {2}.",
                              groupCtr, group.Value, group.Index);
            int captureCtr = 0;
            foreach (Capture capture in group.Captures) {
               captureCtr++;
               Console.WriteLine("      Capture: {0}: '{1}' at position {2}.",
                                 captureCtr, capture.Value, capture.Index);
            }
         }
      }
      Console.WriteLine();

      pattern = @"(a\1|(?(1)\1)){2}";
      Console.WriteLine("Regex pattern: {0}", pattern);
      match = Regex.Match(input, pattern);
         Console.WriteLine("Matched '{0}' at position {1}.",
                           match.Value, match.Index);
      if (match.Groups.Count > 1) {
         for (int groupCtr = 1; groupCtr <= match.Groups.Count - 1; groupCtr++)
         {
            Group group = match.Groups[groupCtr];
            Console.WriteLine("   Group: {0}: '{1}' at position {2}.",
                              groupCtr, group.Value, group.Index);
            int captureCtr = 0;
            foreach (Capture capture in group.Captures) {
               captureCtr++;
               Console.WriteLine("      Capture: {0}: '{1}' at position {2}.",
                                 captureCtr, capture.Value, capture.Index);
            }
         }
      }
   }
}
// The example displays the following output:
//       Regex pattern: (a\1|(?(1)\1)){0,2}
//       Match: '' at position 0.
//          Group: 1: '' at position 0.
//             Capture: 1: '' at position 0.
//
//       Regex pattern: (a\1|(?(1)\1)){2}
//       Matched 'a' at position 0.
//          Group: 1: 'a' at position 0.
//             Capture: 1: '' at position 0.
//             Capture: 2: 'a' at position 0.
Imports System.Text.RegularExpressions

Module Example
    Public Sub Main()
        Dim pattern, input As String

        pattern = "(a\1|(?(1)\1)){0,2}"
        input = "aaabbb"

        Console.WriteLine("Regex pattern: {0}", pattern)
        Dim match As Match = Regex.Match(input, pattern)
        Console.WriteLine("Match: '{0}' at position {1}.",
                          match.Value, match.Index)
        If match.Groups.Count > 1 Then
            For groupCtr As Integer = 1 To match.Groups.Count - 1
                Dim group As Group = match.Groups(groupCtr)
                Console.WriteLine("   Group: {0}: '{1}' at position {2}.",
                                  groupCtr, group.Value, group.Index)
                Dim captureCtr As Integer = 0
                For Each capture As Capture In group.Captures
                    captureCtr += 1
                    Console.WriteLine("      Capture: {0}: '{1}' at position {2}.",
                                      captureCtr, capture.Value, capture.Index)
                Next
            Next
        End If
        Console.WriteLine()

        pattern = "(a\1|(?(1)\1)){2}"
        Console.WriteLine("Regex pattern: {0}", pattern)
        match = Regex.Match(input, pattern)
        Console.WriteLine("Matched '{0}' at position {1}.",
                          match.Value, match.Index)
        If match.Groups.Count > 1 Then
            For groupCtr As Integer = 1 To match.Groups.Count - 1
                Dim group As Group = match.Groups(groupCtr)
                Console.WriteLine("   Group: {0}: '{1}' at position {2}.",
                                  groupCtr, group.Value, group.Index)
                Dim captureCtr As Integer = 0
                For Each capture As Capture In group.Captures
                    captureCtr += 1
                    Console.WriteLine("      Capture: {0}: '{1}' at position {2}.",
                                      captureCtr, capture.Value, capture.Index)
                Next
            Next
        End If
    End Sub
End Module
' The example displays the following output:
'       Regex pattern: (a\1|(?(1)\1)){0,2}
'       Match: '' at position 0.
'          Group: 1: '' at position 0.
'             Capture: 1: '' at position 0.
'       
'       Regex pattern: (a\1|(?(1)\1)){2}
'       Matched 'a' at position 0.
'          Group: 1: 'a' at position 0.
'             Capture: 1: '' at position 0.
'             Capture: 2: 'a' at position 0.

See also