Finding "consistency" or multiple observations in a dataset...

[oncearunner]

I have a dataset with multiple observations per person.:

1. For each person in the dataset for more than one year, how many years of the total that we have data for are they showing up?

2. Sort this somehow so I know the people who are showing up most.

[PGStats · Accepted solution]

Try these :

/* Long shape dataset: person identified by ID, numeric variable : X */
proc sql;
create table want as
select ID, count(X) as n
from myLongData

group by ID
order by desc n;

/* Wide shape dataset: person identified by ID, numeric variables : X2001,...,X2011 */
data want(keep=ID n);
set myWideData;
n = n(of X2001-X2011);
run;
proc sort data=want; by n descending; run;

PG

[oncearunner]

Thanks! I could get the long data set code to work. I really appreciate it...