SAS Arrays and DO Loop Made Easy

SAS Arrays : Introduction

It provides a simple, appropriate way to process a group of variables in a SAS DATA step.

Syntax

Array array-name {number-of-elements} list-of-variables;

Note: You can use [ ] or { } or ( ) for defining number of elements in the ARRAY statement.

Examples

1.  ARRAY ABC[5] a b c d e; 

In the example above, ABC is an array-name, 5 implies the number of variables in array and "a b c d e" are the fields that make up the array.

2. ARRAY ABC[*] a b c d e; 

In the example above, SAS would automatically calculate the number of variables in array.

3. ARRAY ABC[*] X1-X10;

Where the X1 variable contains the X1 value, X2 contains the X2 value, etc.

4. ARRAY ABC[*]  $ X1-X10;

If the variables are of character type then use $ sign before specifying list of variables.

Sample Data

SAS Array : Example

data temp;
input x1 x2 x3 x4$ x5$;
cards;
1 2 3 AA BB
2 3 4 AB CC
3 4 5 AC DD
4 5 6 AD EE
5 6 7 AE FF
6 7 8 AF GG
;
run;

Example I : Numeric variables having value greater than 3 need to be replaced with missing value

data test;
set temp;
array nvars {3} x1-x3;
do i = 1 to 3;
if nvars{i} > 3 then nvars{i} =.;
end;
run;  

Output : Array Statement

Why i is 4 in the output data set?

The first time the loop processes, the value of count is 1; the second time, 2; and the third time, 3. At the beginning of the fourth iteration, the value of count is 4, which is found to be greater than the stop value of 3 so the loop stops. However, the value of i is now 4 and not 3, the last value before it would be greater than 3 as the stop value.

Note : We can drop variable "i" with drop statement or drop data set option.

Improvised version of the above code 

data test;
set temp;
array nvars (*) _numeric_;
do i = 1 to dim(nvars);
if nvars{i} > 3 then nvars{i} =.;
end;
drop i;
run;

Notes - 

1. The "_numeric_" is used to specify all the numeric variables.
2. The DIM function returns the number of elements (variables).

Example II. : Extract first letter of all the character variables

data test;
set temp;
array cvars (*) _character_;
do i = 1 to dim(cvars);
cvars{i} = substr(cvars{i},1,1);
end;
drop i;
run;

 Note - The "_character_" is used to specify all the character variables.

Example III. : Extract first letter and fill in the new character variables

data test;
set temp;
array cvars (*) _character_;
array dvars (*) $ x6 X7;
do i = 1 to dim(cvars);
dvars{i} = substr(cvars{i},1,1) ;
end;
drop i;
run;

Example IV : Assign Initial Values in a SAS Array

data abcd;
set temp;
array nvars (*) _numeric_;
array pvars (*) px1 px2 px3;
array pctinc {3} _temporary_ (1.1 , 1.2 ,1.3); do i = 1 to dim(nvars);
pvars{i} = nvars{i} * pctinc{i};
end;
drop i;
run;  

Notes -

1. In the above example, we are multiplying variables' values with different numbers.
2. When the key word _TEMPORARY_ is used in a ARRAY statement, data elements are created but are not stored in the data file.

Example V : Calculate Percentage Growth

data abcd;
set temp;
array nvars(*) _numeric_;
array diff{2} _temporary_;
array percent{2};
do i = 1 to 2;
diff{i} = nvars{i +1} - nvars{i};
percent{i} = diff{i} / nvars{i} ;
end;
drop i;
run;

Using the OF Operator in a SAS Array


The following two codes are equivalent :

array gnp (*) x y z;
sumgnp = sum(of gnp(*));

OR

sumgnp = sum(x,y,z); 

*Calculate the mean;
  mean_score = mean(of gnp(*));

* Calculate the minimum;
   min_score = min(of gnp(*));

Suppose you are asked to create a flag in cases wherein sum of variables x1,x2 and x3 is greater than 10.

data test;
set temp;
array nvars (*) x1-x3;
if sum(of nvars(*)) > 10 then flag =1;
else flag=0;
run;

DO OVER LOOP

The DO OVER loop is one of the most useful DO loops. It can be used with an array when indexing of the array is not needed.

data test;
set temp;
array nvars _numeric_;
do over nvars;
if nvars > 3 then nvars = .;
end;
run;

SAS Tutorials : 100 Free SAS Tutorials

Spread the Word!
Share Share Tweet

Related Posts

About Author:

Deepanshu founded ListenData with a simple objective - Make analytics easy to understand and follow. He has over 10 years of experience in data science. During his tenure, he has worked with global clients in various domains like Banking, Insurance, Private Equity, Telecom and Human Resource.

While I love having friends who agree, I only learn from those who don't
Let's Get Connected Email LinkedIn