Ruby helper for parsing dates of an unknown format.

parse_date.rb

# ParseDate
#
# A helper for parsing dates of an unknown format.
#
# Takes a randomly formatted date and makes a best guess at the date, or throws and exception if
# there's no best guess.  Will take international dates into account if you pass in a 'short form' date
# to help suggest a starting point for the search.
# Will also try Chronic parsing to get relative dates (yesterday, tomorrow, in 3 days, etc.)
#
# Examples:
#   date = ParseDate.parse('08-09-1977')
#   date = ParseDate.parse('09-08-1977', :short_form => '%d/%m/%y')
#   date = ParseDate.parse('August 09, 1977')
#   date = ParseDate.parse('yesterday')
#
# ParseDate is free and open... you can redistribute it and/or modify
# it anyway you see fit.  No warrantee or guarantee provided :)

require 'date'
require 'rubygems'
require 'chronic'
require 'activesupport'


module ParseDate
  
  # Parse a date of unknown format.
  # Optionally, you can pass in a "short_form" to suggest potential characteristics expected.
  # It will be used to determine if the date is international format, or US formatted
  # (Note: Date separators don't matter with short dates, they will be stripped out)
  def self.parse( date_string, options={} )
    date = nil
    short_form = options[:short_form]
    
    if(day_before_month?(short_form))
      # Normalize separators
      shorted_date_string = date_string.gsub(/[\.-]/,"/")
      shorted_short_form = short_form.gsub(/[\.-]/,"/")
      
      if(long_year?(date_string))
        date ||= Date.strptime(shorted_date_string, shorted_short_form.gsub(/\%y/,"\%Y")) rescue nil
      else
        date ||= Date.strptime(shorted_date_string, shorted_short_form.gsub(/\%Y/, "\%y")) rescue nil
      end
    end
    date ||= Chronic.parse(date_string).to_date rescue nil
    date ||= Date.parse(date_string) # allow this exception through if we can't parse at all
    date
  end
  
  private
  
  # determine if this is an international date format (day before month before year)
  def self.day_before_month?(short_form)
    return false if short_form.nil?
    (/\%d/i =~ short_form) < (/\%m/i =~ short_form)
  end
  
  # Figure out if we have a long (4 digit) year
  def self.long_year?(date)
    # if no spaces...
    if(date.strip.index(' ').nil?)
      # split on -, /, or . (valid separators for dates I guess.)
      parts = date.split(/[-\/\.]/)
      # if the first bit, or last bit if 4 chars
      (parts.first.size == 4) || (parts.last.size == 4)
    end
  end
  
end

parse_date_test.rb

# Tests for ParseDate
# Drop this in you with your unit tests.
#
# Lots of examples of the formats that are accepted.

require File.dirname(__FILE__) + '/../test_helper'
require 'parse_date'

class ParseDateTest < ActiveSupport::TestCase
  
  def test_simple_dates
    # Simple Date
    assert_date '1977-08-19'
    assert_date '19/08/1977'
    assert_date '08-19-1977'
    assert_date '08-09-1977', :expect=>'1977-08-09'
  end
  
  def test_simple_date_with_short_form
    # Simple Date
    assert_date '1977-08-19', :short_form=>'%m/%d/%y'
    assert_date '19/08/1977', :short_form=>'%m/%d/%y'
    assert_date '08-19-1977', :short_form=>'%m/%d/%y'
    assert_date '08-09-1977', :short_form=>'%m/%d/%y', :expect=>'1977-08-09'
  end

  def test_simple_date_short_year
    # Simple Date - short year
    assert_date '19/08/77'
    assert_date '08-19-77'
    assert_date '08-09-77', :expect=>'1977-08-09'
  end
  
  def test_long_form
    # Long Form
    assert_date 'August 19, 1977'
    assert_date 'August 9, 1977', :expect=>'1977-08-09'
  end

  def test_chronic_strings
    # Chonic Relative Date
    assert_date 'yesterday', :expect => (Date.today - 1.day)
    assert_date 'tomorrow', :expect => (Date.today + 1.day)
  end
  def test_international_dates_long_year
    # Date Flipped - Long Year
    assert_date '09/08/1977', :short_form=>'%d/%m/%y', :expect=>'1977-08-09'
    assert_date '09-08-1977', :short_form=>'%d/%m/%y', :expect=>'1977-08-09'
    assert_date '09.08.1977', :short_form=>'%d/%m/%y', :expect=>'1977-08-09'
    assert_date '09/08/1977', :short_form=>'%d/%m/%Y', :expect=>'1977-08-09'
    assert_date '09-08-1977', :short_form=>'%d/%m/%Y', :expect=>'1977-08-09'
    assert_date '09.08.1977', :short_form=>'%d/%m/%Y', :expect=>'1977-08-09'
  end
  
  def test_international_dates_short_year
    # Date Flipped - Short Year
    assert_date '09/08/77', :short_form=>'%d/%m/%y', :expect=>'1977-08-09'
    assert_date '09-08-77', :short_form=>'%d/%m/%y', :expect=>'1977-08-09'
    assert_date '09.08.77', :short_form=>'%d/%m/%y', :expect=>'1977-08-09'
    assert_date '09/08/77', :short_form=>'%d/%m/%Y', :expect=>'1977-08-09'
    assert_date '09-08-77', :short_form=>'%d/%m/%Y', :expect=>'1977-08-09'
    assert_date '09.08.77', :short_form=>'%d/%m/%Y', :expect=>'1977-08-09'
  end
  
  def test_international_dates_short_year_month_day
    # Date Flipped - Short Year - Sort Month/Day
    assert_date '9/8/77', :short_form=>'%d/%m/%y', :expect=>'1977-08-09'
    assert_date '9-8-77', :short_form=>'%d/%m/%y', :expect=>'1977-08-09'
    assert_date '9.8.77', :short_form=>'%d/%m/%y', :expect=>'1977-08-09'
    assert_date '9/8/77', :short_form=>'%d/%m/%Y', :expect=>'1977-08-09'
    assert_date '9-8-77', :short_form=>'%d/%m/%Y', :expect=>'1977-08-09'
    assert_date '9.8.77', :short_form=>'%d/%m/%Y', :expect=>'1977-08-09'
  end
  
  private
  def assert_date(input, options={})
    options[:expect] ||= '1977-08-19'
    expect = options[:expect]
    expect = Date.parse(expect) if expect.is_a?(String)
    got = nil
    
    got = ParseDate.parse(input, :short_form => options[:short_form])
    assert_equal(expect, got)
  end
  
end