Lesson 4, Part 3: SAS Functions, Arrays, and Loops (Additional Code Examples)

The SAS code snippets below calculate the overall weighted percentage total for a course 4197 student (1 observation) using the fictitious data.

Step 1: Create a SAS data set named HAVE with the following variables:

• ID (character variable)
• TEST1 to TEST5 (5 numeric variables containing test scores)
• ASSIGNMENT1, MIDTERM, FINAL (3 additional numeric variables containing the respective assessment scores)
• The DATALINES; ... ; code block inputs just one record.

ods html close;
options nodate nonotes nosource;
data Have;
input id $ TEST1-TEST5 ASSIGNMENT1 MIDTERM FINAL;
datalines;
X1 50 30 100 70 50 72 40 80
;

13 The SAS System

NOTE: Writing HTML5(SASPY_INTERNAL) Body file: _TOMODS1
14 The SAS System

E3969440A681A2408885998500000007

Step 2: Create a second data set (HAVE2) based on HAVE. The data step does the following:

• The CALL SORTN routine sorts the values of the TEST1-TEST5 values in the descending order (largest to smallest). This routine sorts variables within a single observation (row). It does not change the row order.

  As a sightlight, PROC SORT (not used here) sorts an entire dataset by one or more variables and thus modifies the order of the dataset itself.

• Defines three arrays:

  • raw[7] stores the 7 numeric variables (TEST1-TEST4, ASSIGNMENT1, MIDTERM, and FINAL) after dropping the lowest test score among the TEST1 through TEST5 variables .

  • weight[7] assigns predefined weights (temporary array, not stored in the dataset).

  • wp[7] holds the weighted values for each corresponding variable.

  • The DO loop loops through each element in raw, multiplies it by its
    corresponding weight, and stores the result in wp (i.e., applying the same calculation to multiple variables).

  • wpt is the weighted total (sum(OF P:) sums all variables that starts with P).

  The DROP statement drops the loop variable i (not needed in the final dataset). Applies weights to TEST1-TEST4, ASSIGNMENT1, MIDTERM, and FINAL.

options nocenter nodate nosource;
data have2;
 set have;
     call sortN(test5, test4, test3, test2, test1); 
	 array raw[7] TEST1-TEST4 ASSIGNMENT1  MIDTERM FINAL; 
     array weight[7] _temporary_ (.05, .05, .05, .05 
				                   ,.10,.35,.35);
	 array wp[7]  P_TEST1-P_TEST4 P_ASSIGNMENT1 
               P_MIDTERM P_FINAL; 
     do i = 1 to 7;
       wp[i] = raw[i]*weight[i];
     end;
      wpt=sum(OF P:);
 drop i;
 run;

17 The SAS System

E3969440A681A2408885998500000009

The PROC PRINT step below lists the observations from the HAVE2 dataset without observation numbers (noobs), displaying:

• id
• All variables starting with P_ (P_TEST1 to P_FINAL), containing the weighted value
• wpt (overall weighted percentage total)

The FORMAT statement formats wpt to 2 decimal places.

proc print data=work.have2 noobs; 
var id P: wpt; 
Format wpt 6.2;
run;

The SAS System

id	P_TEST1	P_TEST2	P_TEST3	P_TEST4	P_ASSIGNMENT1	P_MIDTERM	P_FINAL	wpt
X1	5	3.5	2.5	2.5	7.2	14	28	62.70

Use Cases of SAS Arrays

SAS arrays allow you to process multiple variables efficiently in a DATA step. They are commonly used for loop-based operations, reducing repetitive code and improving maintainability.

data training;
 miles = 30;
 do i = 1 to 6;
  miles + miles*.04;
   output;
 end;
 run;

proc print;
run;

The SAS System

Obs	miles	i
1	31.2000	1
2	32.4480	2
3	33.7459	3
4	35.0958	4
5	36.4996	5
6	37.9596	6

data training;
 miles = 30;
 do i = 1 to 6;
  miles + miles*.04;
   *output;
 end;
 run;

proc print;
run;

The SAS System

Obs	miles	i
1	37.9596	7

data investment;
 do year = 1 to 5;
  invest+1000;
    do month=1 to 12 by 3;
      invest+50;
      output;
    end;
 end;
 run;

proc print;
run;

The SAS System

Obs	year	invest	month
1	1	1050	1
2	1	1100	4
3	1	1150	7
4	1	1200	10
5	2	2250	1
6	2	2300	4
7	2	2350	7
8	2	2400	10
9	3	3450	1
10	3	3500	4
11	3	3550	7
12	3	3600	10
13	4	4650	1
14	4	4700	4
15	4	4750	7
16	4	4800	10
17	5	5850	1
18	5	5900	4
19	5	5950	7
20	5	6000	10