Re: [Boost-users] [DateTime] Daily Schedules

Subject: Re: [Boost-users] [DateTime] Daily Schedules
From: John Rocha (jrr_at_[hidden])
Date: 2011-01-05 11:41:54

Thank you for the tips on the interval library.

Unfortunately this doesn't achieve what i want. This will deliver the "same" 
results as the contains() member from the time_period class. The underlying 
problem is that the reference points are specific instances of time. A ptime 
isn't just say one o'clock (1:00). A ptime is actually 1:00 on a specific day.

So one could easily make a time range for a single day. I was hoping for an easy 
way to have a time range for any day, or for a range of days (8:00 - 5:00 M-F).

I have ideas on how to extend the classes to achieve this. However, I was hoping 
this already existed.

Still, I appreciate the pointer to the interval class/library, I can envision 
usages for this in other areas of our code already.

Regards,

-=John

On 12/17/2010 3:51 PM, Joachim Faulhaber wrote:
> Hi John,
>
> 2010/12/17 John Rocha<jrr_at_[hidden]>:
>> I was looking at the DateTime library for a scheduling module in my
>> application. I was very excited to see time periods (time_period) and its
>> method contains().
>>
>> I could set up a time_period for say 8:00-17:00, and then when something
>> happened I could create a ptime and see if the time_period contains my
>> ptime. (i.e. did my action happen between 8:00 and 17:00.)
>>
>> However, I realized that a time period is much more specific than I want. A
>> time_period is tied to a specific date, i.e. 8:00-17:00 on Sep. 20 1999.
>>
>> I read the documentation, and I didn't see (or missed) anything that might
>> be what I would call an hour period -- a range of time that is independent
>> of the date. I don't think it's a duration, because I understood that to
>> mean the amount of time between two time points (i.e. +/- 5 hours).
>>
>> I have some ideas on how I might extend this to achieve what I want, but
>> before going that route, I wanted to double check if others have encountered
>> this and if so how they resolved it. Was it done with a bit of the library
>> that I'm overlooking? Or did you have to extend the class to achieve the
>> desired results?
> probably not. But maybe intervals and interval containers might be
> handy for scheduling problems as Jeff has pointed out. There is a code
> snippet that might be helpful:
>
> ---------------------------------------
> #include<boost/icl/ptime.hpp>
> #include<boost/icl/interval.hpp>
>
> using namespace std;
> using namespace boost;
> using namespace boost::icl;
> using namespace boost::posix_time;
>
> void time_test()
> {
>      ptime t_a = time_from_string("2010-12-24 19:30");
>      ptime t_b = time_from_string("2010-12-24 20:30");
>      time_duration a_2_b = t_b - t_a;
>      cout<<  "Duration of ["<<  t_a<<  ","<<  t_b<<  ") ="
>           <<  a_2_b<<  endl;
>
>      interval<ptime>::type a_b
>          = interval<ptime>::right_open(t_a, t_b);
>      cout<<  "Duration of "<<  a_b<<  " ="
>           <<  icl::size(a_b)<<  endl;
>      time_duration half = a_2_b / 2;
>      ptime t_m = t_a + half;
>
>      cout<<  a_b<<  (icl::contains(a_b, t_m) ?
>                      " contains " : "does not contain ")
>           <<  t_m<<  endl;
> }
>
> Boost.Icl is a new library on intervals and interval containers that
> will ship with the next release 1.46. Meanwhile you can find download
> information here:
> http://www.joachim-faulhaber.de
>
> Maybe you might find some examples about date and time processing helpful:
> http://www.joachim-faulhaber.de/boost_icl/doc/libs/icl/doc/html/boost_icl/examples/party.html
> http://www.joachim-faulhaber.de/boost_icl/doc/libs/icl/doc/html/boost_icl/examples/time_grids_for_months_and_weeks.html
> http://www.joachim-faulhaber.de/boost_icl/doc/libs/icl/doc/html/boost_icl/examples/man_power.html
>
> HTH
> Joachim
>

-- 
2 much txting mks u 1 bad splr!!!!