Select the first, second, third variable based on the data

[Smitha9]

Hi,

I have a dataset:

ID    Sub_ID   Measure_year         Ht        Wt

1       12             06/06/1995            .           .

1       13             07/08/1997            .        135

1       14             07/29/1987          66          .

1       15             08/01/1988          56       140

I want the dataset:

where I need to select the Ht which has the first value in it based on the earliest Measure_year with values in Ht and Wt.

ID          Sub_ID        Measure_time           Ht     Wt

1              14               07/29/1987              66     135

Can I do this with more ID's?

thank in advance.

-Smitha

[Tom · Accepted solution]

Not sure what you are after, but you might just want to use the UPDATE statement for this.  Just use the same dataset with zero observations as the BASE dataset to treat all of the observations as TRANSACTIONS.

data want;
  update have(obs=0) have;
  by id;
run;

If you want the FIRST instead of the LAST then sort the input dataset by ID DESCENDING MEASURE_YEAR 

You could also try using the MERGE statement.

data ht;
  set have;
  where not missing(ht);
  by id;
  if first.id;
  keep id measure_year ht ;
  rename measure_year=ht_year;

run;
data wt;
  set have;
  where not missing(wt);
  by id;
  if first.id;
  keep id measure_year wt ;
  rename measure_year=wt_year;
run;

data want;
  merge ht wt;
  by id;
run;