Date:         Thu, 8 Feb 2007 14:42:10 -0500
Reply-To:     hurley.gj@PG.COM
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "George J. Hurley" <hurley.gj@PG.COM>
Subject:      Re: Sum days
Comments: To: Syb it <sas_datalover@YAHOO.COM>
Content-Type: text/plain; charset="iso-8859-1"

Try something this:

 data days_late;
 input id date_pick_check:date7. inc_days @@;
 format date_pick_check date7.;
 cards;
 1 01-Jan-00 31 1 03-Feb-00 28 1 04-Mar-00 31 1 27-Mar-00 30 1 01-May-00 
31
 2 01-Mar-01 31
 3 01-Jun-03 30 3 10-Jul-03 31 3 01-Aug-03 30 3 01-Sep-03 31
 4 01-Apr-99 30 4 05-May-99 31 4 01-Jun-99 30
 ;
 run;
 proc print; run;

proc sort data=days_lateb; by id date_pick_check; run;

data days_lateb; set days_late; by id;
k=date_pick_check+0;
y=lag(date_pick_check);
z=lag(inc_days);
if not first.id then do;
diff=date_pick_check-y;
end;
late=diff-z;
run;

proc sql;
select id, sum(late) as totlate from days_lateb where late>0 group by id;
quit;


George J. Hurley
The Procter and Gamble Company
Miami Valley Innovation Center, Box 30
11810 East Miami River Road
Cincinnati, OH 45252
P: (513) 627-0692
F: (513) 945-3441




Syb it <sas_datalover@YAHOO.COM>
Sent by: "SAS(r) Discussion" <SAS-L@listserv.uga.edu>
02/08/2007 02:22 PM
Please respond to Syb it

 
        To:     SAS-L@listserv.uga.edu
        cc: 
        Subject:        Re: [SAS-L] Sum days


  Dear All,
                                                   date_
                                                   pick_
                                  Obs    id      check    inc_days

                                    1     1    01JAN00       31
                                    2     1    03FEB00       28
                                    3     1    04MAR00       31
                                    4     1    27MAR00       30
                                    5     1    01MAY00       31
                                    6     2    01MAR01       31
                                    7     3    01JUN03       30
                                    8     3    10JUL03       31
                                    9     3    01AUG03       30
                                   10     3    01SEP03       31
                                   11     4    01APR99       30
                                   12     4    05MAY99       31
                                   13     4    01JUN99       30
  Questions:

  I am trying to figure out on average how late a person is to pick up 
their check. Consider person # 1 for example:
  (a) five checks were picked at given dates
  (b) Check is supposed to last a given month ($1/day)
  They picked up their 1st check on 01jan00, then 2nd on Feb 3rd. This 
person was late to pick up their second check by 2 days because each 
dollar lasts a day (that's the assumption). Now I need to sum for each 
person how late they would have been considering all check they would have 

picked up (all records).
  NB: Checks can be collected earlier before the previous one is depleted 
(for example on 27 March for Id=1) was an early pick up.

  Please help with some code ;)

  data days_late;
  input id date_pick_check:date7. inc_days @@;
  format date_pick_check date7.;
  cards;
  1 01-Jan-00 31 1 03-Feb-00 28 1 04-Mar-00 31 1 27-Mar-00 30 1 01-May-00 
31
  2 01-Mar-01 31
  3 01-Jun-03 30 3 10-Jul-03 31 3 01-Aug-03 30 3 01-Sep-03 31
  4 01-Apr-99 30 4 05-May-99 31 4 01-Jun-99 30
  ;
  run;
  proc print; run;


---------------------------------
The fish are biting.
 Get more visitors on your site using Yahoo! Search Marketing.