14.3 正则表达式应用

14.3.4 查找嵌套标记

 

代码

     Sub  FindNestedHtmlTags()
         '  Find all nested HTML tags in a file. (eg. <table>...</table>)
         Dim  text  As   String   =  File.ReadAllText( " test.htm " )
         Dim  re  As   New  Regex( " <(?<tag>(table|tr|td|div|span))[\s>] " , RegexOptions.IgnoreCase)
         For   Each  m  As  Match  In  re.Matches(text)
             '  We've found an open tag. Lets look for open & close versions of this tag.
             Dim  tag  As   String   =  m.Groups( " tag " ).Value
             Dim  openTags  As   Integer   =   1
             Dim  pattern2  As   String   =   String .Format( " ((?<open><{0})[\s>]|(?<close>)</{0}>) " , tag)
             Dim  found  As   String   =   Nothing
             Dim  re2  As   New  Regex(pattern2, RegexOptions.IgnoreCase)

             For   Each  m2  As  Match  In  re2.Matches(text, m.Index  +   1 )
                 If  m2.Groups( " open " ).Success  Then
                    openTags  +=   1
                 ElseIf  m2.Groups( " close " ).Success  Then
                    openTags  -=   1
                     If  openTags  =   0   Then
                        found  =  text.Substring(m.Index, m2.Index  +  m.Length  +   1   -  m.Index)
                         Exit   For
                     End   If
                 End   If
             Next
             '  Display this match.
             If  found  IsNot   Nothing   Then
                Console.WriteLine(found)
                Console.WriteLine( " ------------------------------------------ " )
             Else
                Console.WriteLine( " Unmatched tag {0} at index {1} " , tag, m.Index)
             End   If
         Next
     End Sub

 

 

14.3.5 分析数据文件

  

代码

Imports  System.Text.RegularExpressions
Imports  System.IO

Module  ParsingDataFiles
     Sub  ParsingFixedWidthDataFiles()
         Dim  text  As   String   =
             " John  Smith   New York "   &  vbCrLf  &
             " Ann   Doe     Los Angeles "   &  vbCrLf
         Dim  seps()  As   String   =  {vbCrLf}
         Dim  lines()  As   String   =  text.Split(seps, StringSplitOptions.RemoveEmptyEntries)

         Dim  pattern  As   String   =   " ^(?<first>.{6})(?<last>.{8})(?<city>.+)$ "
         Dim  re  As   New  Regex(pattern)
         For   Each  line  As   String   In  lines
             Dim  m  As  Match  =  re.Match(line)
            Console.WriteLine( " First={0}, Last={1}, City={2} " ,
                              m.Groups( " first " ).Value.TrimEnd(),
                              m.Groups( " last " ).Value.TrimEnd(),
                              m.Groups( " city " ).Value.TrimEnd())
         Next
     End Sub

     ' 分析不带引号的字符串
     Sub  ParsingDelimitedDataFiles()
         Dim  text  As   String   =
             " John , Smith, New York "   &  vbCrLf  &
             " Ann, Doe, Los Angeles "   &  vbCrLf
         Dim  seps()  As   String   =  {vbCrLf}
         Dim  lines()  As   String   =  text.Split(seps, StringSplitOptions.RemoveEmptyEntries)

         Dim  pattern  As   String   =   " ^\s*(?<first>.*?)\s*,\s*(?<last>.*?)\s*,\s*(?<city>.*?)\s*$ "
         Dim  re  As   New  Regex(pattern)
         For   Each  line  As   String   In  lines
             Dim  m  As  Match  =  re.Match(line)
            Console.WriteLine( " First={0}, Last={1}, City={2} " ,
                              m.Groups( " first " ).Value.TrimEnd(),
                              m.Groups( " last " ).Value.TrimEnd(),
                              m.Groups( " city " ).Value.TrimEnd())
         Next
     End Sub

     ' 分析带引号的字符串
     Sub  ParsingDelimitedQuotedDataFiles()
         Dim  text  As   String   =
             " 'John, P.' , ""Smith"" , ""New York"" "   &  vbCrLf  &
             " 'Robert ""Slim""', """" , ""Los Angeles, CA"" "   &  vbCrLf
         Dim  seps()  As   String   =  {vbCrLf}
         Dim  lines()  As   String   =  text.Split(seps, StringSplitOptions.RemoveEmptyEntries)

         Dim  pattern  As   String   =
             " ^\s*(?<q1>[""'])(?<first>.*?)\k<q1>\s*,\s*(?<q2>[""'])(?<last>.*?)\k<q2>\s*,\s*(?<q3>[""'])(?<city>.*?)\k<q3>\s*$ "

         Dim  re  As   New  Regex(pattern)
         For   Each  line  As   String   In  lines
             Dim  m  As  Match  =  re.Match(line)
            Console.WriteLine( " First={0}, Last={1}, City={2} " ,
                              m.Groups( " first " ).Value.TrimEnd(),
                              m.Groups( " last " ).Value.TrimEnd(),
                              m.Groups( " city " ).Value.TrimEnd())
         Next
     End Sub

     ' 只有当该字段包含有分隔符字符时才将文本字段放在引号内
     Sub  ParsingDelimitedQuotedDataFiles2()
         Dim  text  As   String   =
             " ""John, P."" , Doe , ""Los Angeles, CA"" "   &  vbCrLf  &
             " Robert, Smith, New York "   &  vbCrLf
         Dim  seps()  As   String   =  {vbCrLf}
         Dim  lines()  As   String   =  text.Split(seps, StringSplitOptions.RemoveEmptyEntries)

         Dim  pattern  As   String   =
             " ^\s*(?<q1>[""']?)(?<first>.*?)(?(q1)\k<q1>)\s*,\s*(?<q2>[""']?)(?<last>.*?)(?(q2)\k<q2>)\s*,\s*(?<q3>[""']?)(?<city>.*?)(?(q3)\k<q3>)\s*$ "

         Dim  re  As   New  Regex(pattern)
         For   Each  line  As   String   In  lines
             Dim  m  As  Match  =  re.Match(line)
            Console.WriteLine( " First={0}, Last={1}, City={2} " ,
                              m.Groups( " first " ).Value.TrimEnd(),
                              m.Groups( " last " ).Value.TrimEnd(),
                              m.Groups( " city " ).Value.TrimEnd())
         Next
     End Sub
End Module

 

 

14.3.6 分析和计算表达式

 

14.3.7 分析代码

 

14.3.8 使用正则表达式

 

代码

     Sub  PlayPoker()
        PlayPokerHand( " StraightFlush " ,  " 1H " ,  " 2H " ,  " 3H " ,  " 4H " ,  " 5H " )
        PlayPokerHand( " StraightFlush " ,  " 3H " ,  " 4H " ,  " 5H " ,  " 6H " ,  " 7H " )
        PlayPokerHand( " StraightFlush " ,  " 6C " ,  " 7C " ,  " 8C " ,  " 9C " ,  " TC " )
        PlayPokerHand( " StraightFlush " ,  " 7S " ,  " 8S " ,  " 9S " ,  " TS " ,  " JS " )
        PlayPokerHand( " StraightFlush " ,  " 8S " ,  " 9S " ,  " TS " ,  " JS " ,  " QS " )
        PlayPokerHand( " StraightFlush " ,  " TD " ,  " JD " ,  " QD " ,  " KD " ,  " 1D " )

        PlayPokerHand( " Flush " ,  " 8S " ,  " JS " ,  " QS " ,  " KS " ,  " 1S " )
        PlayPokerHand( " Flush " ,  " 7H " ,  " 8H " ,  " 9H " ,  " TH " ,  " KH " )
        PlayPokerHand( " Flush " ,  " 8C " ,  " 9C " ,  " TC " ,  " QC " ,  " KC " )

        PlayPokerHand( " FourOfAKind " ,  " 8C " ,  " 9D " ,  " 8D " ,  " 8H " ,  " 8S " )
        PlayPokerHand( " FourOfAKind " ,  " TS " ,  " TC " ,  " QD " ,  " TD " ,  " TH " )
        PlayPokerHand( " FourOfAKind " ,  " 1C " ,  " 1S " ,  " 8S " ,  " 1H " ,  " 1D " )

        PlayPokerHand( " FullHouse " ,  " 1C " ,  " 1D " ,  " 8C " ,  " 1D " ,  " 8S " )
        PlayPokerHand( " FullHouse " ,  " 9C " ,  " 9S " ,  " 8D " ,  " 8S " ,  " 9H " )
        PlayPokerHand( " FullHouse " ,  " QC " ,  " TD " ,  " QD " ,  " TS " ,  " QS " )

        PlayPokerHand( " Straight " ,  " 3S " ,  " 7S " ,  " 5S " ,  " 4S " ,  " 6D " )
        PlayPokerHand( " Straight " ,  " 6D " ,  " 7S " ,  " 8H " ,  " 9H " ,  " TD " )
        PlayPokerHand( " Straight " ,  " 7S " ,  " 8D " ,  " 9D " ,  " TH " ,  " JC " )
        PlayPokerHand( " Straight " ,  " TC " ,  " JS " ,  " QS " ,  " KS " ,  " 1S " )

        PlayPokerHand( " ThreeOfAKind " ,  " 9C " ,  " 9S " ,  " 8H " ,  " TD " ,  " 9D " )
        PlayPokerHand( " ThreeOfAKind " ,  " TC " ,  " TS " ,  " 7S " ,  " TH " ,  " 1D " )
        PlayPokerHand( " ThreeOfAKind " ,  " TC " ,  " TS " ,  " 8S " ,  " TH " ,  " KH " )

        PlayPokerHand( " TwoPairs " ,  " 1C " ,  " 1F " ,  " 8F " ,  " DP " ,  " 8Q " )
        PlayPokerHand( " TwoPairs " ,  " 9C " ,  " KF " ,  " 8F " ,  " 8P " ,  " 9Q " )
        PlayPokerHand( " TwoPairs " ,  " 1C " ,  " DF " ,  " QF " ,  " DP " ,  " QQ " )

        PlayPokerHand( " OnePair " ,  " 1C " ,  " 1H " ,  " KH " ,  " TS " ,  " 8S " )
        PlayPokerHand( " OnePair " ,  " 9C " ,  " KH " ,  " QH " ,  " 8S " ,  " 9D " )
        PlayPokerHand( " OnePair " ,  " 1S " ,  " TS " ,  " QC " ,  " 8C " ,  " QC " )

        PlayPokerHand( " HighCard " ,  " 1C " ,  " QC " ,  " KC " ,  " TD " ,  " 8D " )
        PlayPokerHand( " HighCard " ,  " TC " ,  " KC " ,  " QD " ,  " 8D " ,  " 9H " )
        PlayPokerHand( " HighCard " ,  " 1C " ,  " TC " ,  " KD " ,  " 8D " ,  " QH " )
        PlayPokerHand( " HighCard " ,  " 8D " ,  " QD " ,  " TD " ,  " 9D " ,  " KH " )
     End Sub

     Private   Sub  PlayPokerHand( ByVal  expectedScore  As   String ,  ByVal   ParamArray  cards()  As   String )
         Dim  score  As   String   =  PokerGame.EvalPokerScore2(cards)
        Console.WriteLine( " {0} {1} {2} {3} {4} => {5} " , cards( 0 ), cards( 1 ), cards( 2 ), cards( 3 ), cards( 4 ), score)
         If  score  <>  expectedScore  Then  Console.WriteLine( "     WRONG! (expected {0}) " , expectedScore)
     End Sub