topic Re: selecting cases in SAS Data Management

selecting cases

Cdot — Fri, 27 Nov 2015 14:46:19 GMT

Hi all,

I have a data set with 2 observations per person, in long format, and I want to select people of a certain age at T1 and a certain place of residence at T1, plus their T2 record, where residence may have changed. I can figure out how to select all the T1 records I want, but not the associated T2 record – any help would be much appreciated!

Re: selecting cases

pearsoninst — Fri, 27 Nov 2015 15:11:03 GMT

Can you provide an example dataset?

Re: selecting cases

Cdot — Fri, 27 Nov 2015 15:34:39 GMT

Sure

ID	Household ID	year of interview	Age at interview	residence	etc	etc
1	1	2004	46	1
1	1	2006	48	1
2	1	2004	67	1
2	1	2006	69	2
3	2	2004	78	3
3	2	2006	79	3
4	3	2004	67	1
4	3	2006	89	2

Re: selecting cases

Reeza — Fri, 27 Nov 2015 15:40:26 GMT

Once you have the T1 records, you can use a SQL step to get all records from the original table.
Assuming your source data is called HAVE and the table with records selected is called T1_Selected you can do the following:

proc sql;
create table want as
select *
from have
where ID in (select ID from T1_Selected);
quit;

Re: selecting cases

Cdot — Fri, 27 Nov 2015 20:53:04 GMT

Another very basic question - how do I create the first table?

Re: selecting cases

Cdot — Fri, 27 Nov 2015 20:53:59 GMT

And, each participant is identified by both a personal ID and a household ID.

Thank you!

Re: selecting cases

Patrick — Fri, 27 Nov 2015 23:47:42 GMT

"how do I create the first table?"

The "have" table? So you don't have your table already in SAS? In what format is your source data (excel, csv,...?). What SAS client are you using (SAS EG, SAS Studio?)

"And, each participant is identified by both a personal ID and a household ID."

The data you posted looks like a person is fully identified by "personal ID". "household ID" appears to be what it says: Grouping persons together into households.