You would likely be much better off actually using a real date if you are comparing to a single fixed date.
I think that data step code can be pretty much reduced to:
data &ds.2 ;
set &ds.1;
if 0< mdy(max(&mm,1),max(&dd,1),&yy) le '01NOV2019'd;
run;
The function MDY returns a SAS date value from numeric variables of month, day and year. Actually if the values are character but resolve to numbers then SAS will attempt to use the numeric version. The special value of a quoted string in DATE9 format ddMONyyyy with the d at the end tells SAS that this is a date literal and can compare to the date value created by the MDY function.
If any of the month, day and year values are missing or result in an invalid date such as 31 November, the function MDY will return a missing value.
The use of Max with the &mm and &dd variables means that if one of the variables is missing then the month 1 or day 1 will be used to create a valid date for comparison. Which basically replicates the choices you made for the different ways of comparing with missing values.
If you really need that OUT variable to indicate the date compared then use:
data &ds.2 ;
set &ds.1;
&out = mdy(max(&mm,1),max(&dd,1),&yy);
format &out yymmdd10.;
if 0< &out le '01NOV2019'd;
run;
An IF without any "then" clause is a subsetting if and will only keep the observations where the If is true. Which is why I have the "if 0< " part. Missing values are always less than any given value so you want to exclude those as well which would occur with a missing &yy variable.
Caution: if you have invalid date combinations like &mm > 12 or an &dd value not valid for a given month (and year in the case of February and leap years) then the mdy function will return missing.
Click image to register for webinar
