SAS Tips

Licence

Confirm expiry date using proc setinit; run;

Details are written to the log

To update your SAS licence run the license tool from the SAS windows start menu 

ASSIGN LIBRARY

Libname  <libref>  <path>;

Libref max 8 characters, either letters, underscore, or numbers but cannot start with number; no characters.

It is good to start a programme with path macro:  %let path=<folder path>;

To invoke the macro: libname <libref> "&path";

MACRO VARIABLES REQUIRE DOUBLE QUOTES

To change a libref: libname <libref> clear;

DATA STEP

Data <new name>; 

Retain <variables>; can be used to set order of variables (must come before set statement)

Set <libref.file>;

Length <variable><$> W or <variable> w.d; where W=number of print places for character variable or for numerical W=bites and .d = decimal places

Assign new variables ; type new variable name and then define

Label;

Format

Length;

Where; create a data subset by limiting to a particular level of a variable

Drop <variables>; if drop statement is used within set statement then variables are not read into the PDV (drop= ) and are not available for processing, otherwise they can be read in any statement from pdv

Keep <variables>; limits the variables that are saved to the new dataset

If condition;

Multiple data sets an be created in one data step, which is best done with a select function:

Select (<variable>);

When ('level1') output data1;

When ('level2') output data2;

Otherwise; optional to otherwise output other

End;

To avoid errors due to capitalisation, the upcase function can be inserted eg select (upcase(<variable>))

Do-end can be added to a select group to execute multiple statements

When ('level1') do;

Newvariable=1; to new data set 'newvariable' is created with value 1

Output data1;

End;

To collapse levels of a variable, e.g, if few observations in one cateory:

Define new variable to = old variable; follow with if statement

<new_variable> = <old_variable>;

if old_variable=3 then new_variable=2; this collapses 3 into 2.

Convert character to numerical variable:

numvar = INPUT(charvar, best32.);

comp_score = comp_score + 0

PROC SORT

Used to create sorted report or merge two data sets.

To avoid permanently writing to data set use out option in proc statement:

proc sort data=< . > out=<name temporary data set>;

by descending <variable> or ascending <variable> ; list variables in order for sorting

PROC CONTENTS

Used to show descriptor portion of data set:

proc contents data=< . > <options>;

Can also be used to display contents of a library:

Proc contents data=<libref>._all_ <options>;

Options

PROC PRINT

Creates reports

Proc print data=< . > <options>;

Var <variables>; prints only listed variables

Sum <variable>; sums numerical variables

ID <numeric ID variable>; replaces the observation number with custom ID

Format <variable><$> <format command >; invokes a predefined format

Where;

Options

FORMATTING

Global formatting

Title<n> "text of title";  where n=line

This is a global statement; to turn off insert title statement with no text;

Footnote uses same syntax as title

A programme can have up to 10 title rows and 10 footnotes rows in use at any one time

To display current date & time use the following macro statements at beginning of programme:

%let timenow=%sysfunc (time(), time.)

%let datenow=%sysfunc (date(), date9.)

Insert &timenow &datenow in title or footnote statement

Label statement

Used in data step to rename a variable

Label <variable>=' '  <variable>=' ' ;

Using the split=' ' option to invoke the label in a proc print step.

Formats

Invoke a SAS or user defined format in a data step or proc print step with the format statement

Format  <variable> <format name>.; format name must include $ prefix if the variable is a character variable, the dot after the format name allows SAS to recognise the format

Format statements have the form: <variable><$> format<w>.<d> {where $ = character variable, w=width, d=decimal}

User defined formats can be defined using proc format:

Proc format;

Value <$><format name> 'AU'='Australia' 'US'='United States' other='miscoded'; character format name must begin with $.

Proc format;

Value <format name>   low - 499='low'

 500 - 1000='medium'

1000 - high='high';

SAS formats

Length

If reassigning variable names SAS truncates the character limit to length of first level specified

To change this use a length statement, which must come before the IF./THEN statements : length <variable> <$> w.d;

CONDITIONAL PROCESSING

IF <argument> {restricts data set to the condition; IF cannot be used in a procedure - used WHERE}

IF-THEN/ELSE

Use ELSE IF for mutually exclusive statements (more efficient)

Only 1 executable statement is allowed per IF-THEN statement; to perform multiple arguments then use IF-DO statement:

If <argument> then do; argument 1; Argument 2; argument …; end;

DO / END

If …. Then do; arguments…..; end;

Else do….; end; run;

IN operator  can be used to pick up equivalent character variables eg if x in('y', 'Y') then y='yes' means that y=yes if x is y or Y

Numerical operators

Logical  operators

Modify where expressions: not / and  / or

Eg where city not in ('London', 'Rome','Paris') = city is not London, Rome or Paris

Comparison operators

Eg where country in ('AU','US') = country either AU or US / where order_type in (1,2,3)  = order type 1,2 or 3

NB inclusive range for IF statement can be achieved  with: 5 le <variable> le 7 {gives observations between 5-7 inclusive}

Where statements

Used in data and proc print steps; can only be used once in a step and only with permanent variables.

Special where operators (can only be used for where)

? Can be used instead of contains

Like operator: '%N' = includes anything with N; 'T_m' = any word that begins with capital T and ends with small m;

FUNCTIONS

Dates & Time

SAS dates are calculated as days from 1Jan 1960

SAS times are calculated in seconds from midnight on a given day

Date functions

Arithmetic operators

Misc operators

**NB: if one of the variables has a missing value, arithmetic + will return a missing value but sum will return the value of the second variable**

Data cleaning functions

CLEANING

Use upcase or lowcase function to convert all levels of a variable to single case style

Eg country=upcase(country) makes all countries names uppercase

In function can be used to combine incorrect variable levels

Eg if sex in('male', 'M', 'Male') then sex='m'

OUTPUT

Create a data file, R click on the file in SAS explorer tab and view in excel.

To change output display select tools>options>preferences>results

To create reports: ods pdf file="<path>filename.ext " style=<style definition>; <sas code to generate report>;run; ods pdf close; {if more than one ODS statement used then use ods _all_ close} 

Common styles: printer, journal, ocean, rtf, sasweb

if missing(age) then adult=.;

if age > 18 and age < 40 then agegroup = 1;

if 18 < age < 40 then agegroup = 1;

* (multiply), and ** (exponentiate

The array is a shorthand way of referring to a group of variables. In effect,

it provides aliases for them so that each variable can be referred to by using

the name of the array and its position within the array in braces. For example,

q12 could be referred to as qall{12} or, when the variable i has the

value 12, as qall{i}. However, the array only lasts for the duration of the

data step in which it is defined.

Suppose we have 20 variables,

q1 to q20, for which ‘not applicable’ has been coded –1 and we wish to set

those to missing values, we might do it as follows:

array qall {20} q1-q20;

do i= 1 to 20;

if qall{i}=-1 then qall{i}=.;

end;

It may be necessary to delete observations from the data set—for example, if

they contain errors. Deleting erroneous observations is best done by using

the if then statement with the delete statement:

if weightloss > startweight then delete;

In a case like this, it would also be useful to write out a message giving more

information about the observation that contains the error:

if weightloss > startweight then do;

put 'Error in weight data' idno= startweight= weightloss=;

delete;

end;

The put statement writes text (in quotes) and the values of variables to

the log.