Force zeros into multiple fields

[Etoo12121]

Hi,

Using the sample dataset below, what is the best way to make zone1,zone2,zone3 and zone4 to all be zero if the location field is absent? if location field is present, then zone fields should remain as is. I'm trying to do this without writing a long if-then statement.

Patient	Visit	Location	Zone1	Zone2	Zone3	Zone4
1	Week1	Absent
1	Week2	Absent
1	Week3	Present	1		0.5	1.5
2	Week1	Absent
2	Week2	Absent
2	Week3	Present	2	1		0.3
3	Week1	Present		1.5	0.9
3	Week2	Absent
3	Week3	Present	3	0.8	0	0
4	Week1	Present		0.9		0.6
4	Week2	Present	3	1.2	2
4	Week3	Present	3		2

[ballardw · Accepted solution]

When you have to do the same thing to multiple variables on a single observation then the typical tool would be an ARRAY.

Perhaps something like:

data want;
   set have;
   array z (*) zone1 - zone4 ;
   if location = 'Absent' then do i=1 to dim(z);
       z[i] = 0;
   end;
   drop i;
run;

Array statement creates a short cut way to allow accessing related variables. The Z is the name of the array and cannot be an existing variable, the (*) says that the size will be based on the list of variables provided and the zone1-zone4 is the list and expects there to be other variables named zone2 and zone3, sequentially numbered.

The Dim function returns how many elements are in the array so the loop executes once for each variable (in this case), setting each value to 0.

[Etoo12121]

Thank you. Arrays are perfect for my situation

[ChrisNZ]

For so few variables just code it and keep the code simple.

data WANT;
  set HAVE;
  if LOCATION = 'Absent' then do;
    ZONE1=0; ZONE2=0; ZONE3=0; ZONE4=0; 
  end;
run;

[Etoo12121]

this works as well but unfortunately, I have more variables than the 4 zones I listed in the sample. I'll keep this in my back pocket. Thank you

[andreas_lds]

Do you need the zeros in the dataset or in a report? If the later is the case, you could just use option missing="0", before creating the report.

[Etoo12121]

I needed them in the dataset