Prev: Looking for Stockist and Distributors for our fast moving Nutraceutical and Baby food products
Next: PROC TABULATE question
From: data _null_; on 29 Apr 2010 13:22
On Apr 29, 11:46 am, arun <aruku...(a)gmail.com> wrote:
> Hi all,
>
> here is the situation Plz help.
>
> I will be getting the data with  2 weeks or 3 weeks data from sql such
> as
>
> variant            date
> 1                    04/12
> 2                    04/13
> 3                    04/14
> .
>
> .
> .
> continously till the end of the two weeks then i need to seperate the
> data into seperate weeks and send them to diff outputs
>
> i am using this
>
> proc sort data = aru.a
> by variant week;
> run;
>
> data aru.1a;
> set aru.a;
> by variant week
> tempdate = lag(week);
> if first.variant then tempdata = .;
> retain flag;
> if first.variant then flag = 1;
> if week ne tempdate then flag = flag +1;
> run;
>
> select;
> when  1>=flag>=7 and _n_ ne 0 then output sar_1;
> when  8>=flag>=14 and _n_ ne 0 then  output sar_2;
> when  15>=flag>=21 and _n_ ne 0 then  output sar_3;
> when  22>=flag>=28 and _n_ ne 0 then output sar_4;
> when  29>=flag>=35 and _n_ ne 0 then output sar_5;
> otherwise;
> end;
>
> Now the issue is there are some situations where I will be missing
> some days for the weeks in between such as
> from 04/12 it will jump to 04/14 where it will take the data of the
> day of next week to this week
>
> now is there any way we could see that if it misses then it should
> take the output of only 6 days to the dataset like first checking for
> the first 7 consecutive days then execute or any other way.
>
> I don't want to run the code manually 2 - 3 times so I am creating a
> macro with do loop at the bottom so that it can run.
>
> Please help me on this if possible.

Create a new variable "WEEK" using INTNX function to group the dates.
This example uses Monday as the first day of the week. 'WEEK.2'

data test;
input variant date:mmddyy.;
format date mmddyy.;
week = intnx('week.2',date,0);
format week mmddyy.;
cards;
1 04/11/2010
1 04/12/2010
2 04/13/2010
3 04/14/2010
3 04/15/2010
3 04/16/2010
3 04/17/2010
3 04/18/2010
3 04/19/2010
3 04/20/2010
;;;;
run;
proc print;
run;
 | 
Pages: 1
Prev: Looking for Stockist and Distributors for our fast moving Nutraceutical and Baby food products
Next: PROC TABULATE question