dbf 0.1.0

data/lib/dbf.rb

@@ -0,0 +1,2 @@
+require 'date'
+require 'dbf/reader'

data/lib/dbf/reader.rb

@@ -0,0 +1,116 @@
+module DBF
+  
+  DBF_HEADER_SIZE = 32
+  FPT_HEADER_SIZE = 512
+  FPT_BLOCK_HEADER_SIZE = 8
+  
+  class Reader
+    
+    attr_reader :field_count
+    attr_reader :fields
+    attr_reader :record_count
+    
+    def initialize(file)
+      @data_file = File.open(file, 'rb')
+      @memo_file = File.open(file.gsub(/dbf$/i, 'fpt'), 'rb') rescue File.open(file.gsub(/dbf$/i, 'FPT'), 'rb') rescue nil
+      get_header_info
+      get_memo_header_info if @memo_file
+      get_field_descriptors
+    end
+    
+    def get_header_info
+      @data_file.rewind
+      @record_count, @header_length, @record_length = @data_file.read(DBF_HEADER_SIZE).unpack('xxxxVvv')
+      @field_count = (@header_length - DBF_HEADER_SIZE + 1) / DBF_HEADER_SIZE
+    end
+    
+    def get_memo_header_info
+      @memo_file.rewind
+      @memo_next_available_block, @memo_block_size = @memo_file.read(FPT_HEADER_SIZE).unpack('Nxxn')
+    end
+    
+    def has_memo_file?
+      @memo_file ? true : false
+    end
+    
+    def get_field_descriptors
+      @fields = Array.new(@field_count) {|i| Field.new(*@data_file.read(32).unpack('a10xax4ss'))}
+    end
+    
+    def field(field_name)
+      @fields.select {|f| f.name == field_name}
+    end
+    
+    def memo(start_block)
+      @memo_file.rewind
+      @memo_file.seek(start_block * @memo_block_size)
+      memo_type, memo_size, memo_string = @memo_file.read(@memo_block_size).unpack("NNa56")
+      if memo_size > @memo_block_size - FPT_BLOCK_HEADER_SIZE
+        memo_string << @memo_file.read(memo_size - @memo_block_size + FPT_BLOCK_HEADER_SIZE)
+      end
+      memo_string
+    end
+    
+    def unpack_string(field)
+      @data_file.read(field.length).unpack("a#{field.length}")
+    end
+    
+    def records
+      seek(0)
+      Array.new(@record_count) {build_record if @data_file.read(1).unpack('c').to_s == '32'}
+    end
+    
+    alias_method :rows, :records
+    
+    def record(index)
+      seek_to_record(index)
+      build_record if @data_file.read(1).unpack('c').to_s == '32'
+    end
+    
+    def build_record
+      record = Record.new
+      @fields.each do |field| 
+        case field.type
+        when 'N' # number
+          record[field.name] = unpack_string(field)[0].to_i
+        when 'D' # date
+          raw = unpack_string(field).to_s.strip
+          record[field.name] = raw.strip.empty? ? nil : Date.new(*raw.match(/([\d]{4})([\d]{2})([\d]{2})/).to_a.slice(1,3).map {|n| n.to_i}) rescue nil
+        when 'M' # memo
+          starting_block = unpack_string(field).first.to_i
+          record[field.name] = starting_block == 0 ? nil : memo(starting_block)
+        when 'L' # logical
+          record[field.name] = unpack_string(field) =~ /(y|t)/i ? true : false
+        else
+          record[field.name] = unpack_string(field).to_s.strip
+        end
+      end
+      record
+    end
+    
+    def seek(offset)
+      @data_file.seek(@header_length + offset)
+    end
+    
+    def seek_to_record(index)
+      seek(@record_length * index)
+    end
+    
+  end
+  
+  class Field
+    attr_accessor :name, :type, :length, :decimal
+
+    def initialize(name, type, length, decimal)
+      self.name, self.type, self.length, self.decimal = name, type, length, decimal
+    end
+
+    def name=(name)
+      @name = name.gsub(/\0/, '')
+    end
+  end
+  
+  class Record < Hash
+  end
+
+end

data/test/read_test.rb

@@ -0,0 +1,65 @@
+$:.unshift(File.dirname(__FILE__) + "/../lib/")
+require 'test/unit'
+require 'dbf'
+
+class ReadTest < Test::Unit::TestCase
+  
+  def setup
+    @dbf = DBF::Reader.new(File.join(File.dirname(__FILE__),'databases', 'foxpro.dbf'))
+  end
+  
+  def test_records
+    assert_kind_of Array, @dbf.records
+    assert_kind_of Array, @dbf.rows
+  end
+  
+  def test_record
+    assert_equal @dbf.record(0), @dbf.records[0]
+    assert_equal @dbf.record(99), @dbf.records[99]
+  end
+  
+  def test_field_count
+    assert_equal @dbf.field_count, @dbf.fields.size
+  end
+  
+  def test_header_info
+    assert_equal 59, @dbf.field_count
+    assert_equal 975, @dbf.record_count
+    assert @dbf.has_memo_file?
+  end
+  
+  def test_character_fields
+    @dbf.records.each do |record| 
+      assert record['NOM'].is_a?(String)
+    end
+  end
+  
+  def test_date_fields
+    @dbf.records.each do |record| 
+      assert record['DATN'].is_a?(Date) || record['DATN'].nil?
+    end
+  end
+  
+  def test_numeric_fields
+    @dbf.records.each do |record|
+      assert record['NF'].is_a?(Fixnum)
+    end
+  end
+  
+  def test_logical_fields
+    # need a test database that has a logical field
+  end
+  
+  def test_memo_fields
+    @dbf.records.each_with_index do |record, index|
+      if [1,3,5].include?(index)
+        assert record['OBSE'].is_a?(String)
+      elsif [2].include?(index)
+        assert record['OBSE'].nil?
+      else
+        assert record['OBSE'].is_a?(String) || record['OBSE'].nil?
+      end
+    end
+  end
+  
+end

metadata

@@ -0,0 +1,52 @@
+--- !ruby/object:Gem::Specification 
+rubygems_version: 0.9.0
+specification_version: 1
+name: dbf
+version: !ruby/object:Gem::Version 
+  version: 0.1.0
+date: 2006-08-01 00:00:00 -07:00
+summary: A library for reading DBase (or XBase, Clipper, Foxpro, etc) database files
+require_paths: 
+- lib
+email: keithm@infused.org
+homepage: 
+rubyforge_project: 
+description: 
+autorequire: 
+default_executable: 
+bindir: bin
+has_rdoc: false
+required_ruby_version: !ruby/object:Gem::Version::Requirement 
+  requirements: 
+  - - ">"
+    - !ruby/object:Gem::Version 
+      version: 0.0.0
+  version: 
+platform: ruby
+signing_key: 
+cert_chain: 
+post_install_message: 
+authors: 
+- Keith Morrison
+files: 
+- lib/dbf
+- lib/dbf.rb
+- lib/dbf/reader.rb
+- test/databases
+- test/read_test.rb
+- test/databases/foxpro.dbf
+- test/databases/foxpro.fpt
+test_files: []
+
+rdoc_options: []
+
+extra_rdoc_files: []
+
+executables: []
+
+extensions: []
+
+requirements: []
+
+dependencies: []
+