From: Ken Bloom <kbloom@...>
Date: 2006-09-06T12:25:06+09:00
Subject: Re: representing a span of time

On Tue, 05 Sep 2006 15:26:41 -0600, ara.t.howard wrote:

> On Wed, 6 Sep 2006, Brad Tilley wrote:
> 
>> Say I have two time objects represented as floats like so:
>>
>> x = 64299600.0
>> y = 1157489583.2798
>>
>> I want to subtract x from y and then represent the difference as years,
>> months, days, hours, minutes and seconds.
>>
>> I don't see how the Time library would do this. Has anyone done
>> something like this? If so, how?
>
[SNIP-CODE]
> 
> harp:~ > ruby a.rb
> a: 2006-09-05 15:26:06.341147 -06:00
> b: 2006-09-05 17:28:06.341147 -06:00
> d.seconds: 7320.0
> d.minutes: 122.0
> d.hours: 2.03333333333333
> d.days: 0.0847222222222222

I think he wants something more like this, where all of the different
options collectively add up to the total time interval (rather than yours
where each different unit of time nonetheless represents the whole span)

class TimeSpan
   # takes input in seconds because that's what Time#to_i returns
   def initialize(seconds)
      @raw_seconds=seconds
      breakdown
   end
   
   attr_reader :seconds,:minutes,:hours,:days
   attr_reader :raw_seconds

   private
   attr_writer :seconds,:minutes,:hours,:days

   # [:x=,y] represents that y is the maximum number of x before
   # carrying over to the next larger unit of time
   FACTORS=[[:seconds=,60],[:minutes=,60],
      [:hours=,24],[:days=,nil]]

   def breakdown
      higherintervals=@raw_seconds
      FACTORS.each do |method, max|
	 higherintervals, thisinterval= \
	    higherintervals.divmod(max) unless max==nil
	 thisinterval=higherintervals if max==nil
	 send(method,thisinterval)
      end
   end
end

irb(main):020:0> a=Time.now.to_i
=> 1157512271
irb(main):021:0> b=Time.now.to_i
=> 1157512296
irb(main):025:0> TimeSpan.new(b-a)
=> #<TimeSpan:0xa7cbfc58 @days=0, @hours=0, @minutes=0,
   @seconds=25, @raw_seconds=25>

Can't do years and months without more complicated math, and a definite
knowledge of what the start date and end date are. But at least this
should be closer.

--Ken Bloom



-- 
Ken Bloom. PhD candidate. Linguistic Cognition Laboratory.
Department of Computer Science. Illinois Institute of Technology.
http://www.iit.edu/~kbloom1/