require 'arjdbc/jdbc/missing_functionality_helper' module ::ArJdbc module Derby def self.column_selector [/derby/i, lambda {|cfg,col| col.extend(::ArJdbc::Derby::Column)}] end def self.monkey_rails unless @already_monkeyd # Needed because Rails is broken wrt to quoting of # some values. Most databases are nice about it, # but not Derby. The real issue is that you can't # compare a CHAR value to a NUMBER column. ::ActiveRecord::Associations::ClassMethods.module_eval do private def select_limited_ids_list(options, join_dependency) connection.select_all( construct_finder_sql_for_association_limiting(options, join_dependency), "#{name} Load IDs For Limited Eager Loading" ).collect { |row| connection.quote(row[primary_key], columns_hash[primary_key]) }.join(", ") end end @already_monkeyd = true end end def self.extended(*args) monkey_rails end def self.included(*args) monkey_rails end module Column def simplified_type(field_type) case field_type when /smallint/i then :boolean when /real/i then :float when /decimal/i then :decimal else super end end # Post process default value from JDBC into a Rails-friendly format (columns{-internal}) def default_value(value) # jdbc returns column default strings with actual single quotes around the value. return $1 if value =~ /^'(.*)'$/ value end end def adapter_name #:nodoc: 'Derby' end def arel2_visitors require 'arel/visitors/derby' {'derby' => ::Arel::Visitors::Derby, 'jdbcderby' => ::Arel::Visitors::Derby} end include ArJdbc::MissingFunctionalityHelper def index_name_length 128 end # Convert the specified column type to a SQL string. # In Derby, the following cannot specify a limit: # - integer # - boolean (smallint) # - timestamp # - date def type_to_sql(type, limit = nil, precision = nil, scale = nil) #:nodoc: return super unless [:integer, :boolean, :timestamp, :date].include? type native = native_database_types[type.to_s.downcase.to_sym] native.is_a?(Hash) ? native[:name] : native end def modify_types(tp) tp[:primary_key] = "int generated by default as identity NOT NULL PRIMARY KEY" tp[:string][:limit] = 256 tp[:integer][:limit] = nil tp[:boolean] = {:name => "smallint"} tp[:timestamp][:limit] = nil tp[:date][:limit] = nil tp end # Override default -- fix case where ActiveRecord passes :default => nil, :null => true def add_column_options!(sql, options) options.delete(:default) if options.has_key?(:default) && options[:default].nil? sql << " DEFAULT #{quote(options.delete(:default))}" if options.has_key?(:default) super end def classes_for_table_name(table) ActiveRecord::Base.send(:subclasses).select {|klass| klass.table_name == table} end # Set the sequence to the max value of the table's column. def reset_sequence!(table, column, sequence = nil) mpk = select_value("SELECT MAX(#{quote_column_name(column)}) FROM #{quote_table_name(table)}") execute("ALTER TABLE #{quote_table_name(table)} ALTER COLUMN #{quote_column_name(column)} RESTART WITH #{mpk.to_i + 1}") end def reset_pk_sequence!(table, pk = nil, sequence = nil) klasses = classes_for_table_name(table) klass = klasses.nil? ? nil : klasses.first pk = klass.primary_key unless klass.nil? if pk && klass.columns_hash[pk].type == :integer reset_sequence!(klass.table_name, pk) end end def remove_index(table_name, options) #:nodoc: execute "DROP INDEX #{index_name(table_name, options)}" end def rename_table(name, new_name) execute "RENAME TABLE #{quote_table_name(name)} TO #{quote_table_name(new_name)}" end AUTO_INC_STMT2 = "SELECT AUTOINCREMENTSTART, AUTOINCREMENTINC, COLUMNNAME, REFERENCEID, COLUMNDEFAULT FROM SYS.SYSCOLUMNS WHERE REFERENCEID = (SELECT T.TABLEID FROM SYS.SYSTABLES T WHERE T.TABLENAME = '%s') AND COLUMNNAME = '%s'" def add_quotes(name) return name unless name %Q{"#{name}"} end def strip_quotes(str) return str unless str return str unless /^(["']).*\1$/ =~ str str[1..-2] end def expand_double_quotes(name) return name unless name && name['"'] name.gsub(/"/,'""') end def auto_increment_stmt(tname, cname) stmt = AUTO_INC_STMT2 % [tname, strip_quotes(cname)] data = execute(stmt).first if data start = data['autoincrementstart'] if start coldef = "" coldef << " GENERATED " << (data['columndefault'].nil? ? "ALWAYS" : "BY DEFAULT ") coldef << "AS IDENTITY (START WITH " coldef << start coldef << ", INCREMENT BY " coldef << data['autoincrementinc'] coldef << ")" return coldef end end "" end def add_column(table_name, column_name, type, options = {}) add_column_sql = "ALTER TABLE #{quote_table_name(table_name)} ADD #{quote_column_name(column_name)} #{type_to_sql(type, options[:limit], options[:precision], options[:scale])}" add_column_options!(add_column_sql, options) execute(add_column_sql) end def execute(sql, name = nil, binds = []) sql = substitute_binds(sql, binds) if sql =~ /\A\s*(UPDATE|INSERT)/i i = sql =~ /\swhere\s/im if i sql[i..-1] = sql[i..-1].gsub(/!=\s*NULL/, 'IS NOT NULL').gsub(/=\sNULL/i, 'IS NULL') end else sql.gsub!(/= NULL/i, 'IS NULL') end super end # SELECT DISTINCT clause for a given set of columns and a given ORDER BY clause. # # Derby requires the ORDER BY columns in the select list for distinct queries, and # requires that the ORDER BY include the distinct column. # # distinct("posts.id", "posts.created_at desc") # # Based on distinct method for PostgreSQL Adapter def distinct(columns, order_by) return "DISTINCT #{columns}" if order_by.blank? # construct a clean list of column names from the ORDER BY clause, removing # any asc/desc modifiers order_columns = order_by.split(',').collect { |s| s.split.first } order_columns.delete_if(&:blank?) order_columns = order_columns.zip((0...order_columns.size).to_a).map { |s,i| "#{s} AS alias_#{i}" } # return a DISTINCT clause that's distinct on the columns we want but includes # all the required columns for the ORDER BY to work properly sql = "DISTINCT #{columns}, #{order_columns * ', '}" sql end SIZEABLE = %w(VARCHAR CLOB BLOB) def structure_dump #:nodoc: definition="" rs = @connection.connection.meta_data.getTables(nil,nil,nil,["TABLE"].to_java(:string)) while rs.next tname = rs.getString(3) definition << "CREATE TABLE #{tname} (\n" rs2 = @connection.connection.meta_data.getColumns(nil,nil,tname,nil) first_col = true while rs2.next col_name = add_quotes(rs2.getString(4)); default = "" d1 = rs2.getString(13) if d1 =~ /^GENERATED_/ default = auto_increment_stmt(tname, col_name) elsif d1 default = " DEFAULT #{d1}" end type = rs2.getString(6) col_size = rs2.getString(7) nulling = (rs2.getString(18) == 'NO' ? " NOT NULL" : "") create_col_string = add_quotes(expand_double_quotes(strip_quotes(col_name))) + " " + type + (SIZEABLE.include?(type) ? "(#{col_size})" : "") + nulling + default if !first_col create_col_string = ",\n #{create_col_string}" else create_col_string = " #{create_col_string}" end definition << create_col_string first_col = false end definition << ");\n\n" end definition end def remove_column(table_name, column_name) execute "ALTER TABLE #{quote_table_name(table_name)} DROP COLUMN #{quote_column_name(column_name)} RESTRICT" end # Notes about changing in Derby: # http://db.apache.org/derby/docs/10.2/ref/rrefsqlj81859.html#rrefsqlj81859__rrefsqlj37860) # # We support changing columns using the strategy outlined in: # https://issues.apache.org/jira/browse/DERBY-1515 # # This feature has not made it into a formal release and is not in Java 6. We will # need to conditionally support this somehow (supposed to arrive for 10.3.0.0) def change_column(table_name, column_name, type, options = {}) # null/not nulling is easy, handle that separately if options.include?(:null) # This seems to only work with 10.2 of Derby if options.delete(:null) == false execute "ALTER TABLE #{quote_table_name(table_name)} ALTER COLUMN #{quote_column_name(column_name)} NOT NULL" else execute "ALTER TABLE #{quote_table_name(table_name)} ALTER COLUMN #{quote_column_name(column_name)} NULL" end end # anything left to do? unless options.empty? begin execute "ALTER TABLE #{quote_table_name(table_name)} ALTER COLUMN #{quote_column_name(column_name)} SET DATA TYPE #{type_to_sql(type, options[:limit])}" rescue transaction do temp_new_column_name = "#{column_name}_newtype" # 1) ALTER TABLE t ADD COLUMN c1_newtype NEWTYPE; add_column table_name, temp_new_column_name, type, options # 2) UPDATE t SET c1_newtype = c1; execute "UPDATE #{quote_table_name(table_name)} SET #{quote_column_name(temp_new_column_name)} = CAST(#{quote_column_name(column_name)} AS #{type_to_sql(type, options[:limit])})" # 3) ALTER TABLE t DROP COLUMN c1; remove_column table_name, column_name # 4) ALTER TABLE t RENAME COLUMN c1_newtype to c1; rename_column table_name, temp_new_column_name, column_name end end end end def rename_column(table_name, column_name, new_column_name) #:nodoc: execute "RENAME COLUMN #{quote_table_name(table_name)}.#{quote_column_name(column_name)} TO #{quote_column_name(new_column_name)}" end def primary_keys(table_name) @connection.primary_keys table_name.to_s.upcase end def columns(table_name, name=nil) @connection.columns_internal(table_name.to_s, name, derby_schema) end def tables @connection.tables(nil, derby_schema) end def recreate_database(db_name) tables.each do |t| drop_table t end end def quote_column_name(name) #:nodoc: %Q{"#{name.to_s.upcase.gsub(/\"/, '""')}"} end def quoted_true '1' end def quoted_false '0' end def add_limit_offset!(sql, options) #:nodoc: if options[:offset] sql << " OFFSET #{options[:offset]} ROWS" end if options[:limit] #ROWS/ROW and FIRST/NEXT mean the same sql << " FETCH FIRST #{options[:limit]} ROWS ONLY" end end private # Derby appears to define schemas using the username def derby_schema if @config.has_key?(:schema) config[:schema] else (@config[:username] && @config[:username].to_s) || '' end end end end