data import;
input
employee : $10.
sales_territory : $10.
order_date: mmddyy10.
ship_date: mmddyy10.;
format order_date ship_date mmddyy10.; *shows dates as dates instead of numbers;
cards;
John East 1/1/2015 1/5/2015
Bob East 1/1/2015 1/4/2015
John West 1/1/2015 1/3/2015
Bob West 1/1/2015 1/2/2015
John North 1/1/2015 1/1/2015
Bob North 1/1/2015 1/6/2015
;
run;
data have;
set import;
date_diff = intck('day',order_date,ship_date);
run;
proc means data=have;
class sales_territory;*Change to the category/categories that you're interested in;
var date_diff;
output out=want
n=cnt
mean=avg;
run; You're on the right track with datepart() Dates are stored as numbers with 1/1/1960 = 1, 1/2/1960 = 2, &c. To see them as dates, you need to format them with something, like mmddyy10., or date9., &c. You can show differences by subtracting them, or using the intck() function. As far as displaying the dimensions and interactions with regard to averages, proc means will probably suit what you're doing.
... View more