topic REURN ALL ROWS FOR AN ID IF ANY ONE ROW MEETS CONDITION FOR TWO OTHER VARIABLES in SAS Programming

REURN ALL ROWS FOR AN ID IF ANY ONE ROW MEETS CONDITION FOR TWO OTHER VARIABLES

d0816 — Wed, 01 Aug 2018 16:58:48 GMT

In the sample dataset below, I want to select all rows for an ID if the begin date with the most recent date value has status W. I am trying to include date condition in the following code, but its not working. Any suggestions?

proc sql;

create table want as Select *

From have

where ID in (select ID From have where status in ('W'));

quit;

sample data
ID	status	begin date	end date
1	W	12/5/2017	1/31/2018
1	W	12/5/2017	1/31/2018
1	W	12/5/2017	1/31/2018
1	J	11/1/2017	12/4/2017
2	F	4/6/2017	4/6/2017
2	W	11/16/2016	4/5/2017
3	C	12/11/2017	2/4/2018
3	C	12/11/2017	2/4/2018
3	W	10/26/2017	12/10/2017
3	W	10/26/2017	12/10/2017
4	W	5/1/2016	5/31/2018
4	W	5/1/2016	5/31/2018
4	W	5/1/2016	5/31/2018
4	W	5/1/2016	5/31/2018
5	C	12/23/2017	2/7/2018
5	W	12/7/2017	12/22/2017
5	W	12/7/2017	12/22/2017
6	W	12/15/2017	1/10/2018
6	W	12/15/2017	1/10/2018
6	W	12/15/2017	1/10/2018
6	F	5/3/2017	12/14/2017
6	F	5/3/2017	12/14/2017
7	S	4/26/2017	4/30/2017
7	S	4/26/2017	4/30/2017
7	W	11/9/2016	4/25/2017
7	W	11/9/2016	4/25/2017
7	W	11/9/2016	4/25/2017
8	F	12/16/2017	3/29/2018
8	F	12/16/2017	3/29/2018
8	W	4/17/2017	12/15/2017
8	W	4/17/2017	12/15/2017
9	W	8/18/2016
9	W	8/18/2016
10	W	12/8/2017
10	J	10/31/2017	12/7/2017
11	W	9/25/2017	2/20/2018

Dataset I want
ID	status	begin date	end date
1	W	12/5/2017	1/31/2018
1	W	12/5/2017	1/31/2018
1	W	12/5/2017	1/31/2018
1	J	11/1/2017	12/4/2017
4	W	5/1/2016	5/31/2018
4	W	5/1/2016	5/31/2018
4	W	5/1/2016	5/31/2018
4	W	5/1/2016	5/31/2018
6	W	12/15/2017	1/10/2018
6	W	12/15/2017	1/10/2018
6	W	12/15/2017	1/10/2018
6	F	5/3/2017	12/14/2017
6	F	5/3/2017	12/14/2017
9	W	8/18/2016
9	W	8/18/2016
10	W	12/8/2017
10	J	10/31/2017	12/7/2017
11	W	9/25/2017	2/20/2018

Re: REURN ALL ROWS FOR AN ID IF ANY ONE ROW MEETS CONDITION FOR TWO OTHER VARIABLES

Astounding — Wed, 01 Aug 2018 17:15:14 GMT

In general, SQL is the wrong tool to use when you want an outcome that depends on the order of the observations. SQL does not guarantee an order. Here's one way that assumes your data is ordered by ID and descending BEGIN_DATE:

data want;

set have;

by id descending begin_date;

if first.id then do;

output_flag = (status="W");

retain output_flag;

end;

if output_flag;

run;

Re: REURN ALL ROWS FOR AN ID IF ANY ONE ROW MEETS CONDITION FOR TWO OTHER VARIABLES

Tom — Wed, 01 Aug 2018 17:26:56 GMT

You need to include your requirement that the W appear on the last date into your query.

create table want as 
select *
from have
where ID in 
 (select ID 
  from have 
  group by id 
  having (max(begin_date)=begin_date)
     and (status in ('W'))
 )
;

Re: REURN ALL ROWS FOR AN ID IF ANY ONE ROW MEETS CONDITION FOR TWO OTHER VARIABLES

d0816 — Wed, 01 Aug 2018 17:57:14 GMT

After adding descending to the code, the code worked.

Thank you so much. Appreciate your help.

Re: REURN ALL ROWS FOR AN ID IF ANY ONE ROW MEETS CONDITION FOR TWO OTHER VARIABLES

d0816 — Wed, 01 Aug 2018 17:59:06 GMT

This code worked too. Thank you. Appreciate your help.