Hi:
It would be helpful to show ALL of the code that you're using and a small sample of your data. Are your variables in the INPUT dataset named _C4_ and _C5_??? Or, are you trying to use the NOPRINT option on the absolute column names that PROC REPORT creates internally??? Generally speaking, you do not use PROC REPORT's absolute column names in a DEFINE statement.
If, for example, your ACROSS variable was REGION and you had this column statement:
[pre]
COLUMN JOBCODE REGION,N;
DEFINE JOBCODE / GROUP;
DEFINE REGION / ACROSS;
[/pre]
If you had 4 regions (A, B, C, D), then the internal absolute column names (as calculated by PROC REPORT would be:
JOBCODE is column 1 on the report (not an ACROSS item)
count column for REGION A = _c2_
count column for REGION B = _c3_
count column for REGION C = _c4_
count column for REGION D = _c5_
On the other hand, if you had this column statement with 2 variables nested under REGION, it would be different:
[pre]
COLUMN JOBCODE REGION,(N SAL);
DEFINE JOBCODE / GROUP;
DEFINE REGION/ ACROSS;
DEFINE N / 'COUNT';
DEFINE SAL / Mean 'AVG SAL';
[/pre]
Then the absolute columns would be assigned differently:
JOBCODE would still be the first column.
_C2_ would be COUNT for REGION A
_C3_ would be SAL for REGION A
_C4_ would be COUNT for REGION B
_C5_ would be SAL for REGION B
_C6_ would be COUNT for REGION C
_C7_ would be SAL for REGION C
_C8_ would be COUNT for REGION D
_C9_ would be SAL for REGION D
And, if you were COMPUTING an item that was under an ACROSS variable, you would potentially have more absolute column names. But let's say, in the above example that you want to USE the N or count for some reason, but hide it on the report. Then in your DEFINE statement you would NOT use the absolute column names to hide the column, you would do this:
[pre]
COLUMN JOBCODE REGION,(N SAL);
DEFINE JOBCODE / GROUP;
DEFINE REGION/ ACROSS;
DEFINE N / 'COUNT' NOPRINT;
DEFINE SAL / Mean 'AVG SAL';
[/pre]
I'm not sure what you mean by "take missing values"??? Are you computing a value that comes out to missing or do you have an ACROSS report item with all missing values??? For example, with the above code snippet, just knowing the report item names isn't enough, it is also necessary to know whether there are observations with missing values for JOBCODE, missing values for REGION and/or missing values for the variables that are nested under REGION -- In order to figure the best approach.
This would be an instance where seeing a small sample or better description of your data would be useful. You might want to consider opening a track with Tech Support, as they can look at your actual data and ALL your code and help you arrive at a solution.
cynthia