What do these asterisks and dots mean in SAS code? - sas

Question

I am trying to understand the following SAS code snippet:
SUM( (list=.)*. , (list=.B)*. , (list=.D)*. , (list=.Q)*. )
where list contains several elements. My understanding so far is that SUM will create a new list as an output of the same length as list, summing the respective elements in the four entries. I guess list=. or list=.B etc maps the elements of list to 1 where the equality is satisfied and to 0 where it is not, but what does it mean for an element to be equal to a dot .? And what kind of quantity can .B be? Also, what does it mean to multiply by a dot, as in variable * .?
EDIT:
Upon further investigation I found out that "." stands for a NaN numeric value in SAS, so I guess list=. produces a new list where all NaN elements in list are set to 1, and all regular numbers are set to 0. But what does it mean to then multiply the resulting new list by .? And what could .B be?

Answer 1

What you posted kind of looks like SAS code, but not sensible SAS code.
SUM(...) in that code is a function call. The function that can take one or more arguments.
= in that code is the equality comparison operator.
* in the code is the multiplication operator.
. and .B and .D and .Q are missing values. The first is the normal missing values and the other three are examples of the other 27 special missing values that SAS supports.
Normal arithmetic, such as X+Y or A*B, will produce a missing result if any of the arguments are missing. But the SUM(,) function ignores the missing values. So the result is only missing if all of the arguments are missing.
The result of a boolean expression, such as LIST=.D , is either 1 (TRUE) or 0 (FALSE). Note that SAS only uses actual Boolean logic, it does not use the Tri-level logic that some other languages use. So X=. is true when the value of X is missing and false when it is not missing (including any of the 27 special missing values).
So far the code begins to look like an attempt to test if LIST is any of those missing values. So something like:
SUM( (list=.) , (list=.B) , (list=.D) , (list=.Q) )
This will be 1 when LIST is any of those four values. And 0 when none of the test is true.
So it is the same test as using the IN operator. Like this
list in (. .B .D .Q)
But by adding the multiplication by a missing value is what makes it nonsense. Because that will always result in a missing result. So you could just replaced the whole SUM() function call with a period to indicate you wanted a missing value no matter what value LIST had.
If you substitute actual numbers for those missing multiplicand values.
SUM( (list=.)*1 , (list=.B)*2 , (list=.D)*3 , (list=.Q)*4 )
Then perhaps it might make some sense. Now the result is a value that is coded 1 when LIST is missing and 2 when LIST is special missing .B, 3 when .D, 4 when .Q and zero otherwise.