anbt-sql-formatter 0.0.1 → 0.0.2

data/bin/anbt-sql-formatter

@@ -1,8 +1,6 @@
 #! /usr/bin/ruby1.8
 # -*- coding: utf-8 -*-
 
-require "pp"
-
 begin
   require "anbt-sql-formatter/formatter"
 rescue LoadError
@@ -44,7 +42,6 @@ def main
   formatter = AnbtSql::Formatter.new(rule)
   result = formatter.format(src)
   puts result
-  #pp result
 end
 
 main()

data/lib/anbt-sql-formatter/coarse-tokenizer.rb

@@ -1,7 +1,5 @@
 # -*- coding: utf-8 -*-
 
-require "pp"
-
 =begin
 エスケープ文字
 =end
@@ -15,7 +13,7 @@ class CoarseToken
   end
 
   def to_s
-    %Q!<#{@_type}>%s</>! % [@string.gsub("\n", "<br>")]
+    @string
   end
 end
 

data/lib/anbt-sql-formatter/exception.rb

@@ -12,8 +12,6 @@ BlancoSqlFormatterException : SQL整形ツールの例外を表します。
 @author sonota (2009-11-xx)
 =end
 
-require "pp"
-
 
 class AnbtSql
   class FormatterException < IOError

data/lib/anbt-sql-formatter/formatter.rb

@@ -1,7 +1,5 @@
 # -*- coding: utf-8 -*-
 
-require "pp"
-
 require "anbt-sql-formatter/rule"
 require "anbt-sql-formatter/parser"
 require "anbt-sql-formatter/exception"
@@ -11,6 +9,8 @@ require "anbt-sql-formatter/helper" # Stack
 class AnbtSql
   class Formatter
 
+    include StringUtil
+
     @rule = nil
 
     def initialize(rule)
@@ -49,7 +49,7 @@ class AnbtSql
       @function_bracket.clear()
       begin
         isSqlEndsWithNewLine = false
-        if sql_str.endsWith("\n")
+        if sql_str.end_with?("\n")
           isSqlEndsWithNewLine = true
         end
         
@@ -103,8 +103,8 @@ class AnbtSql
             tokens[index + 1].string == "+" &&
             tokens[index + 2].string == ")") 
           tokens[index].string = "(+)"
-          tokens.remove(index + 1)
-          tokens.remove(index + 1)
+          ArrayUtil.remove(tokens, index + 1)
+          ArrayUtil.remove(tokens, index + 1)
         end
         index += 1
       end
@@ -115,17 +115,17 @@ class AnbtSql
       prevToken = nil
 
       (tokens.size - 1).downto(1){|index|
-        token     = tokens.get(index)
-        prevToken = tokens.get(index - 1)
+        token     = ArrayUtil.get(tokens, index)
+        prevToken = ArrayUtil.get(tokens, index - 1)
 
         if (token._type == AnbtSql::TokenConstants::SPACE &&
             (prevToken._type == AnbtSql::TokenConstants::SYMBOL ||
              prevToken._type == AnbtSql::TokenConstants::COMMENT))
-          tokens.remove(index)
+          ArrayUtil.remove(tokens, index)
         elsif ((token._type == AnbtSql::TokenConstants::SYMBOL ||
                 token._type == AnbtSql::TokenConstants::COMMENT) &&
                prevToken._type == AnbtSql::TokenConstants::SPACE)
-          tokens.remove(index - 1)
+          ArrayUtil.remove(tokens, index - 1)
         elsif (token._type == AnbtSql::TokenConstants::SPACE)
           token.string = " "
         end
@@ -138,8 +138,8 @@ class AnbtSql
 
       # Length of tokens changes in loop!
       while index < tokens.size
-        prev  = tokens.get(index - 1)
-        token = tokens.get(index    )
+        prev  = ArrayUtil.get(tokens, index - 1)
+        token = ArrayUtil.get(tokens, index    )
 
         if (prev._type  != AnbtSql::TokenConstants::SPACE &&
             token._type != AnbtSql::TokenConstants::SPACE) 
@@ -153,11 +153,11 @@ class AnbtSql
           # 関数名の後ろにはスペースは入れない
           # no space after function name
           if (@rule.function?(prev.string) &&
-              token.string.equals("(")) 
+              token.string == "(")
             index += 1 ; next
           end
           
-          tokens.add(index,
+          ArrayUtil.add(tokens, index,
                      AnbtSql::Token.new(AnbtSql::TokenConstants::SPACE, " ")
                      )
         end
@@ -178,7 +178,7 @@ class AnbtSql
       index = 0
       # Length of tokens changes in loop!
       while index < tokens.size
-        token = tokens.get(index)
+        token = ArrayUtil.get(tokens, index)
         
         if token._type == AnbtSql::TokenConstants::SYMBOL # ****
 
@@ -208,61 +208,61 @@ class AnbtSql
         elsif token._type == AnbtSql::TokenConstants::KEYWORD # ****
 
           # indentを２つ増やし、キーワードの後ろで改行
-          if (token.string.equalsIgnoreCase("DELETE") ||
-              token.string.equalsIgnoreCase("SELECT") ||
-              token.string.equalsIgnoreCase("UPDATE")   )
+          if (equals_ignore_case(token.string, "DELETE") ||
+              equals_ignore_case(token.string, "SELECT") ||
+              equals_ignore_case(token.string, "UPDATE")   )
             indent += 2
             index += insert_return_and_indent(tokens, index + 1, indent, "+2")
           end
 
           # indentを１つ増やし、キーワードの後ろで改行
-          if @rule.kw_plus1_indent_x_nl.any?{ |kw| token.string.equalsIgnoreCase(kw) }
+          if @rule.kw_plus1_indent_x_nl.any?{ |kw| equals_ignore_case(token.string, kw) }
             indent += 1
             index += insert_return_and_indent(tokens, index + 1, indent)
           end
 
           # キーワードの前でindentを１つ減らして改行、キーワードの後ろでindentを戻して改行。
-          if @rule.kw_minus1_indent_nl_x_plus1_indent.any?{ |kw| token.string.equalsIgnoreCase(kw) }
+          if @rule.kw_minus1_indent_nl_x_plus1_indent.any?{ |kw| equals_ignore_case(token.string, kw) }
             index += insert_return_and_indent(tokens, index    , indent - 1)
             index += insert_return_and_indent(tokens, index + 1, indent    )
           end
 
           # キーワードの前でindentを１つ減らして改行、キーワードの後ろでindentを戻して改行。
-          if (token.string.equalsIgnoreCase("VALUES"))
+          if (equals_ignore_case(token.string, "VALUES"))
             indent -= 1
             index += insert_return_and_indent(tokens, index, indent)
           end
 
           # キーワードの前でindentを１つ減らして改行
-          if (token.string.equalsIgnoreCase("END"))
+          if (equals_ignore_case(token.string, "END"))
             indent -= 1
             index += insert_return_and_indent(tokens, index, indent)
           end
 
           # キーワードの前で改行
-          if @rule.kw_nl_x.any?{ |kw| token.string.equalsIgnoreCase(kw) }
+          if @rule.kw_nl_x.any?{ |kw| equals_ignore_case(token.string, kw) }
             index += insert_return_and_indent(tokens, index, indent)
           end
 
           # キーワードの前で改行, インデント+1
-          if @rule.kw_nl_x_plus1_indent.any?{ |kw| token.string.equalsIgnoreCase(kw) }
+          if @rule.kw_nl_x_plus1_indent.any?{ |kw| equals_ignore_case(token.string, kw) }
             index += insert_return_and_indent(tokens, index, indent + 1)
           end
 
           # キーワードの前で改行。indentを強制的に０にする。
-          if (token.string.equalsIgnoreCase("UNION"    ) ||
-              token.string.equalsIgnoreCase("INTERSECT") ||
-              token.string.equalsIgnoreCase("EXCEPT"   )   ) 
+          if (equals_ignore_case(token.string, "UNION"    ) ||
+              equals_ignore_case(token.string, "INTERSECT") ||
+              equals_ignore_case(token.string, "EXCEPT"   )   ) 
             indent -= 2
             index += insert_return_and_indent(tokens, index    , indent)
             index += insert_return_and_indent(tokens, index + 1, indent)
           end
 
-          if token.string.equalsIgnoreCase("BETWEEN")
+          if equals_ignore_case(token.string, "BETWEEN")
             encounterBetween = true
           end
 
-          if token.string.equalsIgnoreCase("AND")
+          if equals_ignore_case(token.string, "AND")
             # BETWEEN のあとのANDは改行しない。
             if not encounterBetween
               index += insert_return_and_indent(tokens, index, indent)
@@ -272,10 +272,10 @@ class AnbtSql
 
         elsif (token._type == AnbtSql::TokenConstants::COMMENT) # ****
 
-          if token.string.startsWith("/*")
+          if token.string.start_with?("/*")
             # マルチラインコメントの後に改行を入れる。
             index += insert_return_and_indent(tokens, index + 1, indent)
-          elsif /^--/ =~ token.string
+          elsif token.string.start_with?("--")
             index += insert_return_and_indent(tokens, index + 1, indent)
           end
         end
@@ -295,21 +295,21 @@ class AnbtSql
       (tokens.size - 1).downto(4).each{|index|
         next if (index >= tokens.size()) 
 
-        t0 = tokens.get(index    )
-        t1 = tokens.get(index - 1)
-        t2 = tokens.get(index - 2)
-        t3 = tokens.get(index - 3)
-        t4 = tokens.get(index - 4)
+        t0 = ArrayUtil.get(tokens, index    )
+        t1 = ArrayUtil.get(tokens, index - 1)
+        t2 = ArrayUtil.get(tokens, index - 2)
+        t3 = ArrayUtil.get(tokens, index - 3)
+        t4 = ArrayUtil.get(tokens, index - 4)
 
-        if (t4.string.     equalsIgnoreCase("(") &&
-            t3.string.trim.equalsIgnoreCase("" ) &&
-            t1.string.trim.equalsIgnoreCase("" ) && 
-            t0.string.     equalsIgnoreCase(")")   )
+        if (equals_ignore_case(t4.string      , "(") &&
+            equals_ignore_case(t3.string.strip, "" ) &&
+            equals_ignore_case(t1.string.strip, "" ) && 
+            equals_ignore_case(t0.string      , ")")   )
           t4.string = t4.string + t2.string + t0.string
-          tokens.remove(index    )
-          tokens.remove(index - 1)
-          tokens.remove(index - 2)
-          tokens.remove(index - 3)
+          ArrayUtil.remove(tokens, index    )
+          ArrayUtil.remove(tokens, index - 1)
+          ArrayUtil.remove(tokens, index - 2)
+          ArrayUtil.remove(tokens, index - 3)
         end
       }
     end
@@ -321,15 +321,15 @@ class AnbtSql
       # SQLの前後に空白があったら削除する。
       # Delete space token at first and last of SQL tokens.
 
-      token = tokens.get(0)
+      token = ArrayUtil.get(tokens, 0)
       if (token._type == AnbtSql::TokenConstants::SPACE)
-        tokens.remove(0)
+        ArrayUtil.remove(tokens, 0)
       end
       return [] if tokens.empty?
       
-      token = tokens.get(tokens.size() - 1)
+      token = ArrayUtil.get(tokens, tokens.size() - 1)
       if token._type == AnbtSql::TokenConstants::SPACE
-        tokens.remove(tokens.size() - 1)
+        ArrayUtil.remove(tokens, tokens.size() - 1)
       end
       return [] if tokens.empty?
 
@@ -361,33 +361,26 @@ class AnbtSql
       begin
         # 挿入する文字列を作成する。
         s = "\n"
-        # もし１つ前にシングルラインコメントがあるなら、改行は不要。
-        prevToken = tokens.get(index - 1)
-
-        if (prevToken._type == AnbtSql::TokenConstants::COMMENT &&
-            prevToken.string.startsWith("--")) 
-          s = ""
-        end
         
         # インデントをつける。
         indent = 0 if indent < 0 ## Java版と異なる
         s += @rule.indent_string * indent
 
         # 前後にすでにスペースがあれば、それを置き換える。
-        token = tokens.get(index)
+        token = ArrayUtil.get(tokens, index)
         if token._type == AnbtSql::TokenConstants::SPACE
           token.string = s
           return 0
         end
 
-        token = tokens.get(index - 1)
+        token = ArrayUtil.get(tokens, index - 1)
         if token._type == AnbtSql::TokenConstants::SPACE
           token.string = s
           return 0
         end
 
         # 前後になければ、新たにスペースを追加する。
-        tokens.add(index,
+        ArrayUtil.add(tokens, index,
                    AnbtSql::Token.new(AnbtSql::TokenConstants::SPACE, s)
                    )
         return 1

data/lib/anbt-sql-formatter/helper.rb

@@ -1,73 +1,59 @@
-require "pp"
+class AnbtSql
+  class Stack
+    include Enumerable
 
-class Stack
-  include Enumerable
-
-  def initialize
-    @arr = []
-  end
-
-  def each
-    @arr.each{|item|
-      yield item
-    }
-  end
-
-  def clear
-    @arr.clear
-  end
-
-  def push(o)
-    @arr.push o
-  end
-
-  def pop
-    @arr.pop
-  end
-end
+    def initialize
+      @arr = []
+    end
 
+    def each
+      @arr.each{|item|
+        yield item
+      }
+    end
 
-class String
-  def endsWith(c)
-    self[-1] == c ? true : false
-  end
+    def clear
+      @arr.clear
+    end
 
-  def startsWith(c)
-    self[0] == c ? true : false
-  end
+    def push(o)
+      @arr.push o
+    end
 
-  def charAt(n)
-    self[n..n]
+    def pop
+      @arr.pop
+    end
   end
 
-  def equals(str)
-    self == str
-  end
+  module StringUtil
+    def char_at(str, n)
+      if n < 0 || str.size - 1 < n
+        raise IndexOutOfBoundsException
+      end
 
-  def equalsIgnoreCase(other)
-    self.upcase == other.upcase
-  end
+      str.slice(n, 1)
+    end
 
-  def trim
-    self.strip
+    def equals_ignore_case(str_a, str_b)
+      str_a.casecmp(str_b) == 0
+    end
   end
-end
 
+  module ArrayUtil
+    def self.remove(ary, n)
+      ary.delete_at n
+    end
 
-class Array
-  def remove(n)
-    self.delete_at n
-  end
+    def self.get(ary, n)
+      if n < 0 || ary.size - 1 < n
+        raise IndexOutOfBoundsException
+      end
 
-  def get(n)
-    if n >= self.size || n <= -1
-      raise IndexOutOfBoundsException
+      ary[n]
     end
 
-    self[n]
-  end
-
-  def add(n,o)
-    self.insert(n,o)
+    def self.add(ary, n, o)
+      ary.insert(n, o)
+    end
   end
 end

data/lib/anbt-sql-formatter/parser.rb

@@ -10,6 +10,8 @@ require "anbt-sql-formatter/coarse-tokenizer"
 class AnbtSql
   class Parser
 
+    include ::AnbtSql::StringUtil
+
     def initialize(rule)
       @rule = rule
 
@@ -26,7 +28,7 @@ class AnbtSql
 
       # ２文字からなる記号。
       # なお、|| は文字列結合にあたります。
-      @two_character_symbol = [ "<>", "<=", ">=", "||" ]
+      @two_character_symbol = [ "<>", "<=", ">=", "||", "!=" ]
     end
 
 
@@ -49,7 +51,7 @@ class AnbtSql
       return false if space?(c)
       return false if digit?(c)
       return false if symbol?(c)
-      
+
       true
     end
 
@@ -64,7 +66,7 @@ class AnbtSql
     # アンダースコアは記号とは扱いません
     # これ以降の文字の扱いは保留
     def symbol?(c)
-      %w(" ? % & ' \( \) | * + , - . / : ; < = > ).include? c
+      %w(" ? % & ' \( \) | * + , - . / : ; < = > !).include? c
       #"
     end
 
@@ -80,35 +82,36 @@ class AnbtSql
       $stderr.puts "next_token #{@pos} <#{@before}> #{@before.length}" if $DEBUG
 
       start_pos = @pos
-      
+
       if @pos >= @before.length
         @pos += 1
         return nil
       end
-      
-      @char = @before.charAt(@pos)
+
+      @char = char_at(@before, @pos)
 
       if space?(@char)
         workString = ""
-        loop { 
+        loop {
           workString += @char
 
-          @char = @before.charAt(@pos+1)
-          if not space?(@char)
-            @pos += 1
-            return AnbtSql::Token.new(AnbtSql::TokenConstants::SPACE,
-                                        workString, start_pos)
+          is_next_char_space = false
+          if @pos + 1 < @before.size &&
+            space?(char_at(@before, @pos+1))
+              is_next_char_space = true
           end
 
-          @pos += 1
-          
-          if @pos >= @before.length()
+          if not is_next_char_space
+            @pos += 1
             return AnbtSql::Token.new(AnbtSql::TokenConstants::SPACE,
-                                        workString, start_pos)
+                                      workString, start_pos)
+          else
+            @pos += 1
+            next
           end
         }
 
-        
+
       elsif @char == ";"
         @pos += 1
         # 2005.07.26 Tosiki Iga セミコロンは終了扱いではないようにする。
@@ -116,66 +119,51 @@ class AnbtSql
                                     ";", start_pos)
 
       elsif digit?(@char)
-        if /(0x[0-9a-fA-F]+)/       =~ @before[@pos..-1] || # hex
-           /(\d+(\.\d+(e-?\d+)?)?)/ =~ @before[@pos..-1]    # float or scientific
+        if /^(0x[0-9a-fA-F]+)/       =~ @before[@pos..-1] || # hex
+           /^(\d+(\.\d+(e-?\d+)?)?)/ =~ @before[@pos..-1]    # integer, float or scientific
           num = $1
           @pos += num.length
           return AnbtSql::Token.new(AnbtSql::TokenConstants::VALUE,
                                     num, start_pos)
+        else
+          raise "must not happen"
         end
 
-        s = ""
-        while (digit?(@char) || @char == '.') 
-          # if (ch == '.') type = Token.REAL
-          s += @char
-          @pos += 1
-
-          if (@pos >= @before.length) 
-            # 長さを超えている場合には処理中断します。
-            break
-          end
-
-          @char = @before.charAt(@pos)
-        end
-        return AnbtSql::Token.new(AnbtSql::TokenConstants::VALUE,
-                                    s, start_pos)
-        
-        
       elsif letter?(@char)
         s = ""
         # 文字列中のドットについては、文字列と一体として考える。
-        while (letter?(@char) || digit?(@char) || @char == '.') 
+        while (letter?(@char) || digit?(@char) || @char == '.')
           s += @char
           @pos += 1
           if (@pos >= @before.length())
             break
           end
 
-          @char = @before.charAt(@pos)
+          @char = char_at(@before, @pos)
         end
 
         if AnbtSql::Constants::SQL_RESERVED_WORDS.map{|w| w.upcase }.include?(s.upcase)
           return AnbtSql::Token.new(AnbtSql::TokenConstants::KEYWORD,
                                       s, start_pos)
         end
-        
+
         return AnbtSql::Token.new(AnbtSql::TokenConstants::NAME,
                                     s, start_pos)
 
       elsif symbol?(@char)
         s = "" + @char
         @pos += 1
-        if (@pos >= @before.length()) 
+        if (@pos >= @before.length())
           return AnbtSql::Token.new(AnbtSql::TokenConstants::SYMBOL,
                                     s, start_pos)
         end
 
         # ２文字の記号かどうか調べる
-        ch2 = @before.charAt(@pos)
+        ch2 = char_at(@before, @pos)
         #for (int i = 0; i < two_character_symbol.length; i++) {
         for i in 0...@two_character_symbol.length
-          if (@two_character_symbol[i].charAt(0) == @char &&
-              @two_character_symbol[i].charAt(1) == ch2)
+          if (char_at(@two_character_symbol[i], 0) == @char &&
+              char_at(@two_character_symbol[i], 1) == ch2)
             @pos += 1
             s += ch2
             break
@@ -209,7 +197,7 @@ class AnbtSql
       pos = 0
       while pos < coarse_tokens.size
         coarse_token = coarse_tokens[pos]
-        
+
         case coarse_token._type
 
         when :quote_single
@@ -255,10 +243,10 @@ class AnbtSql
     ##
     # ２つ以上並んだキーワードは１つのキーワードとみなします。
     #     ["a", " ", "group", " ", "by", " ", "b"]
-    #  => ["a", " ", "group by",         " ", "b"] 
+    #  => ["a", " ", "group by",         " ", "b"]
     def concat_multiwords_keyword(tokens)
       temp_kw_list = @rule.kw_multi_words.map{|kw| kw.split(" ") }
-      
+
       # ワード数が多い順から
       temp_kw_list.sort{ |a, b|
         b.size <=> a.size
@@ -270,7 +258,7 @@ class AnbtSql
           temp_tokens = tokens[index, target_tokens_size].map {|x|
             x.string.sub(/\s+/, " ")
           }
-          
+
           if /#{kw.join(" ")}/i =~ temp_tokens.join
             tokens[index].string = temp_tokens.join
             (target_tokens_size-1).downto(1).each{|c|
@@ -297,28 +285,28 @@ class AnbtSql
       coarse_tokens = CoarseTokenizer.new.tokenize(sql_str)
 
       prepare_tokens(coarse_tokens)
-      
+
       tokens = []
       count = 0
       @token_pos = 0
       loop {
         token = next_token()
-        
+
         if $DEBUG
           pp "=" * 64, count, token, token.class
         end
-        
+
         if token._type == AnbtSql::TokenConstants::END_OF_SQL
           break
         else
           ;
         end
-        
+
         tokens.push token
         count += 1
         @token_pos += 1
       }
-      
+
       concat_multiwords_keyword(tokens)
 
       tokens