_N_ not working

[R_A_G_]

Hello,

I've got another problem, I am using _N_ to have a count for my observations but for some reason it does not give me the correct count

Thanks
R.A.G
my Code:
data class (rename=(skills1-skills&num_skills=attributes1-attributes&num_skills));
set all_skills1;
array skills{&num_skills} skills1-skills&num_skills;
num=%eval(&num_skills);
if _type_=(2**num)-1 and sum(of skills1-skills&num_skills) ne .;

/*Lambda Intercept*/
num=1;
class=_n_;/****************************HOW???*************/
/*Lambda Main Effect*/
drop num _type_ _freq_ num;

run;

[sbb]

What is the value you are getting? Have you included this statement to verify with accuracy:

PUTLOG _ALL_;

Scott Barry
SBBWorks, Inc.

[R_A_G_]

the values I am getting starts from 66 and ends with 81, I have 4 variables with 16 observations.
Thanks

[Tim_SAS]

_N_ counts the number of times the DATA step has iterated. It is not a count of observations.

See the doc about Automatic Variables.

[NickR]

_N_ is the count of observation read from the input dataset and not the observation written to output dataset.

In your case, you are subsetting the data and then assigning the value of _N_ to class variable. How about replacing the statement "class=_n_;" with "class+1;"

[R_A_G_]

thanks, that worked, but why?
thanks again
R.A.G

[Ksharp]

Because you have "if _type_=(2**num)-1 and sum(of skills1-skills&num_skills) ne .; ".


"if _type_=(2**num)-1 and sum(of skills1-skills&num_skills) ne .; " is actually ""if _type_=(2**num)-1 and sum(of skills1-skills&num_skills) ne . then output;"

That means only obs 66 - 81 suited this condition (i.e. _N_ = 66 - 81),only obs 66 - 81 will output dataset.


Ksharp