DATA STEP

[robertrao]

IS THE "COMMENTED" AND THE "NON SCOMMENTED" LINES DOING THE SAME THING IN THE BELOW CODE????

THANKS

data want;

set ints_exts2;

 array ins int:;

 array outs ext:;

 do i = 1 to dim(ins);

 if ins(i)^=. and outs(i)^=. then duration=sum(duration,outs(i)-ins(i)) /*UNCOMMENTED*/

 *if 2=N(ins(i),outs(i)) then duration=sum(duration,outs(i)-ins(i));    /*COMMENTED*/

end

 ;

days = datepart(duration);

time = timepart(duration);

hrs=time/(60*60);

Total_vent_hours =days*24+ time/(60*60);

vent_days=Total_vent_hours/24;

 run;

[ballardw]

Not exactly as the Not missing, ^=., works for character variables while N(variables) requires all of the arguments to be numeric.

If you are dealing with mixed data types you might want CMISS(ins[i], out[i]) = 0

[Reeza]

If you have any other special missing values that also won't give you the same results.

data have;
a=.; b=2; output;
a=1; b=.N; output;
a=2; b=3; output;
a=.; b=.; output;
a=.A; b=.A; output;
run;

data want;
set have;
if a ne . and b ne . then flag=1;
else flag=0;
x=n(a,b);
run;

[robertrao]

Thanks Reeza.

INS and OUTS arrays contain dates only. Can have missing values.

Wanted toi know if in that case both line output the same results??

Thanks

[ballardw]

You could check with:

compare =(ins(i)^=. and outs(i)^=.) = (N(ins(i),outs(i)));

[Reeza]

If you have a very specific scenario, it's faster to test it than ask. In general, no it won't, for your case it may, as long as you can guarantee that it will always stay that way.