New sub-dataset where the values are the top 25% of initial dataset column

[NKormanik]

Thus, filtering away the lowest 75%, based on a particular variable, and leaving just the top 25% in the new sub-dataset.

Can we actually use the "where" clause for this?

Wondering how you folks might handle it:

data want;
set have;
where
top_25_percent_in_column = ??? ;
run;

Perhaps there isn't currently a SAS 'function' for such a purpose?

If you feel SAS is a bit function-lacking, say so.  Can the equivalent be found by using Python or R within SAS?  Never tried using either in SAS, but I understand there are thousands of Python and R functions.

Thoughts appreciated....

[PaigeMiller]

I assume you want the top 25 percentile (rather than top 25%, I don't know what that means).

proc rank data=have groups=4 out=want descending;
    var yourvariablename;
    ranks rank;
run;

You want any record in WANT where variable RANK is equal to zero.

[Reeza]

---

@PaigeMiller wrote:

I assume you want the top 25 percentile (rather than top 25%, I don't know what that means).

proc rank data=have groups=4 out=want (where = (rank=3)) descending;
    var yourvariablename;
    ranks rank;
run;

You want any record in WANT where variable RANK is equal to zero.

---

Note you can modify @PaigeMiller solution with a WHERE statement to have the data filtered. 

Percentiles and 25% of data don't necessarily align though so you should be clear on the differences and if it meets your needs, especially if you have ties in your data. 

[Ksharp · Accepted solution]

Paige already gave you a solution.

Another way is calcualted Q3 firstly and keep the top 25% via it .

data have;
set sashelp.heart;
run;

proc summary data=have;
var weight;
output out=Q3 q3=q3;
run;

data want;
 set have;
 if _n_=1 then set q3(keep=q3);
 if weight>=q3 ;
 drop q3;
run;