Library Guides: An Introductory Guide to SAS: Working with Dataset(s)

Importing Data

Importing an Excel file

Syntax

proc import

datafile = 'file path\filename.xlsx'

out = new_dataset; run;

Example

proc import

datafile = 'c:\temp\sample.xlsx'

out = sample; run;

The above code reads an Excel file called "sample.xlsx" from "c:\temp" on the computer. It also creates a new dataset called "sample" based on the Excel file.

Importing an SPSS file

Syntax

proc import

datafile = 'file path\filename.sav'

out = new_dataset

dbms = SAV replace; run;

Example

proc import

datafile = 'c:\temp\sample.sav'

out = sample

dbms = SAV replace; run;

The above code reads an SPSS file called "sample.sav" from "c:\temp" on the computer. It also creates a new dataset called "sample" based on the SPSS file.

Importing a CSV file

Syntax

proc import

datafile = 'file path\filename.csv'

out = new_dataset

dbms = CSV replace; run;

Example

proc import

datafile = 'c:\temp\sample.csv'

out = sample

dbms = CSV replace; run;

The above code reads an CSV file called "sample.csv" from "c:\temp" on the computer. It also creates a new dataset called "sample" based on the CSV file.

Printing data

Syntax

proc contents data = dataset;

run;

Example

proc contents data = sample;

run;

The above code prints the content of dataset "sample."

Computing New Variables

Creating a new variable by copying an existing variable

Syntax

var_new = var_old

Example

data sample_new;

set sample;

age_new = age;

run;

The above code creates a new variable "age_new" by copying the value from an existing variable "age" in the dataset.

Creating a new variable with a constant value

Syntax

var_new = constant_number

Example

data sample_new;

set sample;

group = 1;

run;

The above code creates a new variable "group" with a constant value 1.

Creating new variables by arithmetic calculations

Arithmetic Operators

Symbol	Definition
**	exponentiation
*	multiplication
/	division
+	addition
-	subtraction

Example

data calculation;

set sample;

ID_new = ID + 1;

age_new = age - 2;

point = total*5;

weight_kg = weight_lb/2.205;

bmi = (weight / (height*height) ) * 703;

run;

Using if-then statement

Comparison and Logical Operators

Definition	Symbol	Alternative
equal to	=	EQ
not equal to	^=	NE
greater than	>	GT
less than	<	LT
greater than or equal to	>=	GE
less than or equal to	<=	LE
equal to one of a list	in	IN
	&	AND
	\|	OR
		NOT

Example

data sample_new;

set sample;

if age >= 18 then adult = 1;

if age < 18 then adult = 0;

if age >= 18 & gender = 1 then group = 1;

run;

Sorting Data

Sorting data by ascending

Syntax

proc sort data = dataset_old

out = dataset_new;

by var_name;

run;

Example

proc sort data = sample

out = sample_ascending;

by age;

run;

The above code sorts the data by age in ascending order.

Sorting data by descending

Syntax

proc sort data = dataset_old

out = dataset_new;

by descending var_name;

run;

Example

proc sort data = sample

out = sample_descending;

by descending age;

run;

The above code sorts the data by age in descending order.

Subsetting and Merging

Subsetting data

1. By variables

Example

data sample_new;

set sample;

keep number ID age;

run;

The above code creates a subset from dataset "sample" only including variables: number, ID, and age.

2. By dropping missing value

data sample_new;

set sample;

if ID = . then delete;

run;

The above code creates a subset from dataset "sample" only including those observations with a valid ID value.

3. By multiple conditions

data sample_new;

set sample;

if age >= 18 & group = 1 then delete;

run;

The above code creates a subset from dataset "sample" only including those observations with age less than 18 and group number is not 1.

Merging data

Syntax

data dataset_new;

merge dataset1 dataset2;

by var;

run;

Example

data sample_new;

merge sample1 sample2;

by id;

run;

The above code merges two datasets "sample1" and "sample2" into a new dataset "sample_new" by the common variable "id". (Remember to sort the merging datasets by the common variable before merging into a new dataset.)

Stacking data

Syntax

data dataset_new;

set dataset1 dataset2;

run;

Example

data sample_new;

set sample1 sample2;

run;

The above code stacks two datasets "sample1" and "sample2" into a new dataset "sample_new."

Summarizing Data

Printing variables

Syntax

proc print data = dataset;

var var_name1 var_name2 var_name3;

run;

Example

proc print data = sample;

var number age id;

run;

The above code prints values of variables number, age, and id from dataset "sample."

Summarizing continuous variables

Syntax

proc means data = dataset_name;

var var_name;

class var_name;

run;

Example 1

proc means data = dataset_name mean median std min max maxdec = 2;

run;

The above code calculates mean, median, standard deviation, minimum value, maximum value of each variable in dataset "sample." Results only provide 2 decimals.

Example 2

proc means data = sample N mean median std min max maxdec = 2;

var age;

class gender;

run;

The above code calculates number of observation, mean, median, standard deviation, minimum value, maximum value of variable age in dataset "sample" by gender. Results only provide 2 decimals.

Summarizing categorical variables

Syntax

proc freq data = dataset_name;

tables var_name1 var_name2 var_name3;

run;

Example 1

proc freq data = sample;

tables race;

run;

The above code creates a frequency table for variable race in dataset "sample." It includes both frequency and percentage of each category of the variable.

Example 2

proc freq data = sample order = freq;

tables race;

run;

The above code creates a frequency table for variable race in dataset "sample." The table is sorted by frequency of each category within the variable.

Analyzing Data

Pearson Correlation

Syntax

proc corr data = dataset_name;

var var_name1;

run;

Example

proc corr data = sample;

var math science read;

run;

Independent t-test

Syntax

proc ttest data = dataset_name;

var var_name1(outcome variable);

class var_name2(independent variable);

run;

Example

proc ttest data = sample;

var write;

class gender;

run;

Chi-Square Test

Syntax

proc freq data = dataset_name;

tables var_row*var_column /chisq;

run;

Example

proc freq data = sample;

tables sports*location /chisq;

run;

One Way ANOVA

Syntax

proc anova data = dataset_name;

class var_name;

model var_name1 = var_name2;

run;

Example

proc anova data = sample;

class school;

model math = school;

run;

Linear Regression

Syntax

proc reg data = dataset_name;

title "analysis_title";

model outcome_var = independent_var;

run;

Example 1 (simple linear regression)

proc reg data = sample;

title "example 1";

model bmi = weight;

run;

Example 2 (multiple linear regression)

proc reg data = sample;

title "example 2";

model bmi = weight height age;

run;