Difference in output when using WHERE statement on DATA line or below SET line

[Angmar]

Hello,

I realised that the subset of data SAS creates when I use the WHERE statement in different lines is different.

For example, I often use:

DATA x (WHERE=(variable=1)

SET y

. . . 

But I recently used

DATA x;

SET y;

WHERE=variable=1;

 . . .

And the outputs were different. Would someone please explain to me why there is a difference?

Thank you.

[mtnbikerjoshua · Accepted solution]

Hi Angmar,

The difference there is that the first example, using where as a dataset option on the data statement, subsets the output dataset after all the processing in the data step has been done. The second example, using the where statement, subsets the input dataset(s) before all of the data processing. So, if the value of "variable" changes at all during the data step, the results of those two examples will be different.

Hope that explains it for you.

Joshua

[Angmar]

Thank you very much! I understand now.

[data_null__]

WHERE=variable=1;

When used in a data step is an assignment statement NOT a WHERE statement.  

Try using valid syntax.

WHERE variable = 1;