Class ActiveRecord::ConnectionAdapters::JdbcTypeConverter
In: lib/active_record/connection_adapters/jdbc_adapter.rb
Parent: Object

I want to use JDBC‘s DatabaseMetaData#getTypeInfo to choose the best native types to use for ActiveRecord‘s Adapter#native_database_types in a database-independent way, but apparently a database driver can return multiple types for a given java.sql.Types constant. So this type converter uses some heuristics to try to pick the best (most common) type to use. It‘s not great, it would be better to just delegate to each database‘s existin AR adapter‘s native_database_types method, but I wanted to try to do this in a way that didn‘t pull in all the other adapters as dependencies. Suggestions appreciated.

Methods

choose_best_types   choose_type   new  

Constants

AR_TO_JDBC_TYPES = { :string => [ lambda {|r| Jdbc::Types::VARCHAR == r['data_type'].to_i}, lambda {|r| r['type_name'] =~ /^varchar/i}, lambda {|r| r['type_name'] =~ /^varchar$/i}, lambda {|r| r['type_name'] =~ /varying/i}], :text => [ lambda {|r| [Jdbc::Types::LONGVARCHAR, Jdbc::Types::CLOB].include?(r['data_type'].to_i)}, lambda {|r| r['type_name'] =~ /^(text|clob)/i}, lambda {|r| r['type_name'] =~ /^character large object$/i}, lambda {|r| r['sql_data_type'] == 2005}], :integer => [ lambda {|r| Jdbc::Types::INTEGER == r['data_type'].to_i}, lambda {|r| r['type_name'] =~ /^integer$/i}, lambda {|r| r['type_name'] =~ /^int4$/i}, lambda {|r| r['type_name'] =~ /^int$/i}], :decimal => [ lambda {|r| Jdbc::Types::DECIMAL == r['data_type'].to_i}, lambda {|r| r['type_name'] =~ /^decimal$/i}, lambda {|r| r['type_name'] =~ /^numeric$/i}, lambda {|r| r['type_name'] =~ /^number$/i}, lambda {|r| r['precision'] == '38'}, lambda {|r| r['data_type'] == '2'}], :float => [ lambda {|r| [Jdbc::Types::FLOAT,Jdbc::Types::DOUBLE, Jdbc::Types::REAL].include?(r['data_type'].to_i)}, lambda {|r| r['data_type'].to_i == Jdbc::Types::REAL}, #Prefer REAL to DOUBLE for Postgresql lambda {|r| r['type_name'] =~ /^float/i}, lambda {|r| r['type_name'] =~ /^double$/i}, lambda {|r| r['type_name'] =~ /^real$/i}, lambda {|r| r['precision'] == '15'}], :datetime => [ lambda {|r| Jdbc::Types::TIMESTAMP == r['data_type'].to_i}, lambda {|r| r['type_name'] =~ /^datetime$/i}, lambda {|r| r['type_name'] =~ /^timestamp$/i}, lambda {|r| r['type_name'] =~ /^date/i}], :timestamp => [ lambda {|r| Jdbc::Types::TIMESTAMP == r['data_type'].to_i}, lambda {|r| r['type_name'] =~ /^timestamp$/i}, lambda {|r| r['type_name'] =~ /^datetime/i}, lambda {|r| r['type_name'] =~ /^date/i}], :time => [ lambda {|r| Jdbc::Types::TIME == r['data_type'].to_i}, lambda {|r| r['type_name'] =~ /^time$/i}, lambda {|r| r['type_name'] =~ /^date/i}], :date => [ lambda {|r| Jdbc::Types::DATE == r['data_type'].to_i}, lambda {|r| r['type_name'] =~ /^date$/i}, lambda {|r| r['type_name'] =~ /^date/i}], :binary => [ lambda {|r| [Jdbc::Types::LONGVARBINARY,Jdbc::Types::BINARY,Jdbc::Types::BLOB].include?(r['data_type'].to_i)}, lambda {|r| r['type_name'] =~ /^blob/i}, lambda {|r| r['type_name'] =~ /sub_type 0$/i}, # For FireBird lambda {|r| r['type_name'] =~ /^varbinary$/i}, # We want this sucker for Mimer lambda {|r| r['type_name'] =~ /^binary$/i}, ], :boolean => [ lambda {|r| [Jdbc::Types::TINYINT].include?(r['data_type'].to_i)}, lambda {|r| r['type_name'] =~ /^bool/i}, lambda {|r| r['data_type'] == '-7'}, lambda {|r| r['type_name'] =~ /^tinyint$/i}, lambda {|r| r['type_name'] =~ /^decimal$/i}], }   The basic ActiveRecord types, mapped to an array of procs that are used to select the best type. The procs are used as selectors in order until there is only one type left. If all the selectors are applied and there is still more than one type, an exception will be raised.

Public Class methods

Public Instance methods

[Validate]