SAS Study Notes – Chapters 1

🖥

The SAS Environment

Windows, panels and editor colour coding

Interface

Key Windows

Editor — write SAS code (stored as ASCII)
Log — run details, errors in red, warnings in orange
Output — analysis results after a run
Explorer/Results — libraries, datasets, tree view
Graph — appears when graphs are produced

Editor Colour Coding

Green — comments
Red — errors
Dark Blue — SAS keywords
Yellow — data values (classic SAS)
Boundary line — separates steps

💡

SAS code is plain ASCII text — write it in any editor and paste into the SAS Editor. Reference files: FIRST.SAS, SECOND.SAS

⚙️

How SAS Works — Three Steps

DATA → PROC → Output

Concept

1. DATA Step

Enter or manipulate data
Creates a SAS dataset
Runs row by row sequentially

2. PROC Step

Calls built-in procedures
MEANS, PRINT, SORT, IMPORT…

3. Output Step

Observes and directs output
Controlled by ODS statements

Minimal working program

DATA MYDATA;
  INPUT ID NAME $ AGE;
  DATALINES;
  1 Alice 30
  2 Bob   25
  ;
RUN;

PROC MEANS DATA=MYDATA;
  VAR AGE;
RUN;

🔑

Tips, Tricks & Common Errors

Rules for writing valid SAS code

Syntax

Syntax Rules

One statement can span multiple lines
Several statements can share one line
SAS is not case-sensitive
Every statement ends with a semicolon ;

Top 3 Errors

#1 — Missing or misplaced semicolon ;
#2 — Missing RUN; statement
#3 — Unmatched quotation marks

🔴

Always check the Log Window after running code. Errors appear in red. Never assume success just because output appeared.

📦

The DATA Statement & Dataset Names

Temporary vs permanent datasets

Data Step

Temporary (one-part name)

Exists only during the current session
Auto-stored in the WORK library
Example: DATA MYDATA;

Permanent (two-part name)

Saved to disk beyond the session
Format: LIBRARY.DATASET
Example: DATA PER.CLASS;

Temporary vs Permanent

/* Temporary */
DATA MYDATA;
  INPUT ID AGE;
  DATALINES;
  1 25
  2 30
  ;
RUN;

/* Permanent — two-part name */
DATA PER.CLASS;
  SET MYDATA;
RUN;

📋

Variable Names & Types

Naming rules and the three SAS variable types

Fundamentals

Naming Rules

1–32 characters, no blanks
Must start with A–Z or underscore _
Can include numbers (not first character)
Case-insensitive
Valid: GENDER, AGE_IN_1999, _OUTCOME
Invalid: AGE IN 2000, 2000Count, AGE-In-2000

Three Variable Types

Numeric — arithmetic calculations, grouping codes
Character ($) — text/string, not used in arithmetic
Date — stored as a SAS numeric value; many formats accepted (10/15/09, JAN052010, etc.)

📥

Four Methods of Reading Data

Freeform, Compact, Column, and Formatted input

Input

Method 1 — Freeform List Input

/* $ marks character variables; values must be blank-separated */
DATA MYDATA;
  INPUT ID $ SBP DBP GENDER $ AGE WT;
  DATALINES;
  1 120 80 F 34 65
  2 130 85 M 45 80
  ;
RUN;

✅ Advantages

Minimal specification
No fixed column positions required
Best for blank-separated data

❌ Disadvantages

Variables must be in INPUT order
Values need ≥ 1 blank separator
Missing values must use .

Method 2 — Compact Method (@@)

/* @@ keeps reading across lines — multiple subjects per line */
DATA MYDATA;
  INPUT ID AGE @@;
  DATALINES;
  1 25 2 30 3 22 4 19
  ;
RUN;

Method 3 — Column Input

/* Specify exact column ranges — blanks don't need to separate values */
DATA MYDATA;
  INPUT ID 1-2  NAME $ 3-12  SBP 13-15  DBP 16-18;
  DATALINES;
  01Alice     12080
  02Bob       13085
  ;
RUN;

✅ Advantages

No blanks needed between fields
Embedded spaces in names work (John Smith)
Character fields up to 200 chars

❌ Disadvantages

Data must be in fixed column positions
Blank fields read as missing
More specs than list input

Method 4 — Formatted Input (@ pointer)

/* @ pointer = column start; INFORMAT = how to read it */
DATA MYDATA;
  INPUT @1 SBP  3.      /* 3-digit number at col 1 */
        @4 DBP  3.      /* 3-digit number at col 4 */
        @7 DOB  DATE9.; /* DDMONYYYY date at col 7 */
  DATALINES;
  12080 12JAN1990
  ;
RUN;

Common SAS INFORMATs

Format	Raw Value	Read As
Comma7.	$40,000	40000
Comma10.2	190,020.22	190020.22
Dollar10.2	$19020.22	19020.22
Date9.	12JAN1999	SAS numeric date value

💡

Reference files: DFREEFORM.SAS, DCOLUMN.SAS, DINFORMAT.SAS

📂

INFILE Statement & Options

Read external ASCII/text files instead of inline DATALINES

Input

Basic INFILE (must appear before INPUT)

DATA MYDATA;
  INFILE "C:\Data\myfile.txt";
  INPUT ID AGE SBP DBP;
RUN;

DLM = ","

Custom delimiter (e.g. comma for CSV)

DSD

Two consecutive delimiters = missing value

MISSOVER

Moves to next line when obs values run short

FIRSTOBS = n

Start reading at row n (skip headers)

OBS = n

Stop reading at row n

CSV file with options

DATA MYDATA;
  INFILE "C:\Data\myfile.csv"
         DLM="," DSD FIRSTOBS=2;
  INPUT ID NAME $ AGE;
RUN;

💡

Reference file: DINPUT.SAS

🗄️

SAS Libraries & Permanent Datasets

Storing data beyond the current session

Libraries

Temporary Dataset

Single-level name: DATA PEOPLE;
Lost when session ends
Stored in WORK library

Permanent Dataset

Two-part name: MYSASLIB.PEOPLE
Saved as .SAS7BDAT file on disk
Writable via full Windows path

Windows path technique

/* Write using full path */
DATA "C:\SASDATA\PEOPLE";
  INPUT ID NAME $ AGE;
  DATALINES;
  1 Alice 30
  ;
RUN;

/* Read back using the same path */
PROC PRINT DATA="C:\SASDATA\PEOPLE"; RUN;

LIBNAME statement

/* Assign a library shortname to a folder */
LIBNAME SAS_LIB "C:\Users\You\Documents\Library";

/* Use two-part names to read/write */
DATA SAS_LIB.MYDATA; SET MYDATA; RUN;

💡

Also create libraries via: Explorer → New Library → Browse → Save. Reference file: WRITE.SAS

📊

PROC IMPORT & Importing Excel/CSV

Read external files (XLS, XLSX, CSV) into SAS datasets

Import

Excel prep checklist

Row 1 = valid SAS variable names
Each column = one consistent variable type
Blank cells → missing value (.)

Key options

OUT = — output SAS dataset
DATAFILE = — path to input file
DBMS = — file type (XLSX, CSV, XLS)
REPLACE — overwrite existing dataset
SHEET = — which worksheet
GETNAMES = — YES if row 1 has names

PROC IMPORT example

PROC IMPORT
  OUT      = SAS_LIB.EXAMPLE
  DATAFILE = "C:\Data\Example.xlsx"
  DBMS     = XLSX
  REPLACE;
  SHEET    = "Sheet1";
  GETNAMES = YES;
RUN;

/* Verify with PROC CONTENTS */
PROC CONTENTS DATA=SAS_LIB.EXAMPLE; RUN;

💡

PROC CONTENTS shows variable names, types and widths — always use after importing to verify.

🏷️

LABEL Statement

Give descriptive names to cryptic variable names

Data Prep

Syntax & example

DATA MYDATA2;
  SET MYDATA;
  LABEL
    SBP    = "Systolic Blood Pressure"
    DBP    = "Diastolic Blood Pressure"
    WT     = "Weight in Kilograms"
    GENDER = "Biological Sex";
RUN;

/* Display labels as column headers in PROC PRINT */
PROC PRINT DATA=MYDATA2 LABEL; RUN;

💡

Single or double quotes both work. Labels up to 256 characters. Reference file: DLABEL.SAS

🔢

Creating New Variables & SAS Functions

Arithmetic operators and built-in functions

Data Prep

Arithmetic Operators

+ Addition - Subtraction
* Multiplication / Division
** Exponentiation
Order: Parentheses → Exponents → × ÷ → + −

Important Functions

INT(x) — integer part
MAX(x1,x2,…) — maximum
MIN(x1,x2,…) — minimum
SUM(x1,x2,…) — sum
ROUND(x,unit) — round
MDY(m,d,y) — create SAS date
INTCK('interval',s,e) — date diff

Creating variables

DATA MYDATA2;
  SET MYDATA;
  AREA      = WIDTH * LENGTH;
  CELSIUS   = (FAHRENHEIT - 32) * (5/9);
  TOTAL     = SUM(SCORE1, SCORE2, SCORE3);
  AGEYEARS  = INTCK('YEAR', DOB, TODAY());
RUN;

💡

Reference file: DCALC.SAS

🔀

IF-THEN-ELSE & Missing Values

Conditional logic and handling invalid data

Data Prep

IF-THEN-ELSE

DATA MYDATA2;
  SET MYDATA;

  IF      SBP <  120 THEN BP_CAT = "Normal";
  ELSE IF SBP <  130 THEN BP_CAT = "Elevated";
  ELSE IF SBP >= 140 THEN BP_CAT = "High";
RUN;

Assigning & flagging missing values

/* Numeric missing = .  (period) */
IF AGE EQ -9 THEN AGE = .;

/* Character missing = "" (two quote marks) */
IF GENDER NE "M" AND GENDER NE "F"
  THEN GENDER = " ";

/* Subset — delete records from temp dataset only */
IF GENDER EQ "M" THEN DELETE;

📌

DELETE only removes records from the temporary dataset — raw data is never affected. Reference files: DCONDITION.SAS, DMISSING.SAS

✂️

DROP & KEEP Statements

Select which variables to retain in a dataset

Data Prep

DROP and KEEP

/* DROP — read all, discard listed variables */
DATA MYDATA_CLEAN;
  INPUT A B C D E F G;
  DROP E F;
  DATALINES;
  /* data here */
  ;
RUN;

/* KEEP — only retain listed variables */
DATA MYDATA_SLIM;
  SET MYDATA;
  KEEP ID AGE GENDER SBP DBP;
RUN;

💡

Use KEEP when retaining few variables; use DROP when excluding just a few. Same outcome — pick whichever is shorter. Reference file: DKEEP.SAS

🔃

PROC SORT

Sort observations ascending or descending

Sorting

DATA =

Input dataset

OUT =

Output dataset (omit to sort in place)

BY var(s)

One or more sort keys (ascending default)

DESCENDING

Place before a variable for descending order

Sorting examples

/* Ascending (default) */
PROC SORT DATA=MYDATA; BY AGE; RUN;

/* Descending */
PROC SORT DATA=MYDATA; BY DESCENDING AGE; RUN;

/* Multiple variables — sort to new dataset */
PROC SORT DATA=MYDATA OUT=SORTED;
  BY GENDER DESCENDING AGE;
RUN;

📌

Must PROC SORT before using a BY statement in PROC MEANS. Reference file: DSORT.SAS

🔗

Appending & Merging Datasets

Stack rows (SET) or join by a common key (MERGE + BY)

Combining

Append (SET)

Stacks rows from multiple datasets
Compatible structures required

Merge (MERGE + BY)

Joins on a common key variable
Both datasets must be sorted by the BY variable first

Append with SET

DATA COMBINED;
  SET DATASET_A DATASET_B;
RUN;

Merge with MERGE + BY

/* Step 1: sort both by key */
PROC SORT DATA=DEMOGRAPHICS; BY ID; RUN;
PROC SORT DATA=LABRESULTS;   BY ID; RUN;

/* Step 2: merge */
DATA MERGED;
  MERGE DEMOGRAPHICS LABRESULTS;
  BY ID;
RUN;

💡

Reference files: DAPPEND1.SAS, DMERGE.SAS, DSUBSET2.SAS

🏷

PROC FORMAT

Relabel coded values for display without altering raw data

Formatting

Numeric format

PROC FORMAT DATA=MYDATA;
  VALUE FMTMARRIED  0 = "NO"
                     1 = "YES";
RUN;

ODS HTML;
PROC PRINT DATA=MYDATA;
  FORMAT MARRIED FMTMARRIED.;
RUN;
ODS HTML CLOSE;

Character format ($)

PROC FORMAT DATA=MYDATA;
  VALUE $FMTGENDER  "F" = "FEMALE"
                     "M" = "MALE";
RUN;
ODS HTML;
PROC PRINT DATA=MYDATA;
  FORMAT GENDER $FMTGENDER.;
RUN;
ODS HTML CLOSE;

💡

Numeric formats have no $; character formats always start with $. Always end the format call with a period: FMTMARRIED. Reference file: DFORMAT2.SAS

🖨

PROC PRINT (Modified)

Control the appearance of printed output

Output

DOUBLE

Double-spaces output rows

LABEL

Shows variable labels as column headers

N = "text"

Heading + total observation count

OBS = "text"

Custom label for observation column

ROUND

Rounds numeric values before printing

Example

ODS HTML;
PROC PRINT DATA=MYDATA
           DOUBLE LABEL
           N="Total observations:"
           OBS="Record #"
           ROUND;
RUN;
ODS HTML CLOSE;

📌

Options go in the PROC statement, after the dataset name. Reference file: APRINT1.SAS

📤

ODS — Output Delivery System

Export output to HTML, PDF, RTF and more

Output

How it works

Wrap code between ODS …; and ODS … CLOSE;
HTML is the default from SAS 9.4 onwards

Destinations

HTML — browser-viewable (default)
PDF — printable document
RTF — Word-compatible
LISTING — classic plain text

ODS template

ODS HTML;
PROC PRINT DATA=MYDATA; RUN;
ODS HTML CLOSE;

ODS PDF;
PROC MEANS DATA=MYDATA; RUN;
ODS PDF CLOSE;

📝

TITLE & FOOTNOTE Statements

Add text to the top or bottom of every output page

Output

Usage

TITLE    "Employee Survey Results 2024";
TITLE2   "Subset: Department A";
FOOTNOTE "Source: HR Database";

PROC PRINT DATA=MYDATA; RUN;

/* Clear after use */
TITLE; FOOTNOTE;

💡

Supports up to TITLE10 / FOOTNOTE10. Declare without text to clear. Reference file: DTITLE.SAS

💬

Comments in SAS Code

Annotate your programs — always appears green in editor

Syntax

Both comment styles

/* Block comment — can span multiple lines */

* Statement comment — ends with semicolon;

DATA MYDATA;
  /* Inline: explaining this step */
  SET RAWDATA;
RUN;

🟢

All commented text displays in green in the SAS editor. If it's not green, your comment syntax is wrong.

📊

PROC MEANS

Descriptive statistics for quantitative variables

Statistics

What it does

Evaluates quantitative/numerical data
Groups analysis via BY or CLASS
Helps spot outliers quickly

Default statistics

N — observations count
MEAN — arithmetic average
STD — standard deviation
MIN / MAX — extremes

Full syntax

PROC MEANS DATA=MYDATA MAXDEC=2
           N MEAN STD MIN MAX MEDIAN RANGE SUM;
  VAR   variable1 variable2;
  BY    GroupVar;     /* must be pre-sorted */
  /* OR */
  CLASS GroupVar;     /* no pre-sort needed */
RUN;

MAXDEC = n

Decimal places in output

N / NMISS

Non-missing / missing obs counts

MEAN / STD

Average / standard deviation

MIN / MAX

Minimum and maximum values

RANGE / SUM

MAX−MIN / total sum

VAR / STDERR

Variance / standard error

MEDIAN

50th percentile

CLM

Confidence limits for mean

BY vs CLASS

/* BY — must sort first */
PROC SORT DATA=MYDATA; BY GENDER; RUN;
PROC MEANS DATA=MYDATA MEAN STD;
  VAR AGE; BY GENDER;
RUN;

/* CLASS — no sorting required */
PROC MEANS DATA=MYDATA MEAN STD;
  VAR AGE; CLASS GENDER;
RUN;

💡

Use CLASS for convenience; BY when data is already sorted. Reference files: AMEANS.1.SAS, AMEANS.2.SAS

📈

PROC UNIVARIATE

Normality testing, outlier detection and distribution plots

Distribution

What it does

Tests whether data is normally distributed
Identifies outliers
Produces histograms, QQ plots, probability plots

Key output metrics

Skewness — symmetry, should be ≈ 0
Kurtosis — shape, should equal 0
Positive kurtosis → more peaked
Negative kurtosis → flatter

Full syntax

PROC UNIVARIATE DATA=MYDATA NORMAL PLOT;
  VAR       variable1;
  HISTOGRAM variable1 / NORMAL(COLOR=RED W=2);
  QQPLOT    variable1;
  PROBPLOT  variable1;
  INSET     MEAN STD / FORMAT=(4.0);
RUN;

NORMAL

Normality test table (Shapiro-Wilk etc.)

PLOT

Stem-and-leaf, box & probability plots

TRIMMED

Trim extremes before mean calculation

CIBASIC

Confidence intervals for mean, std, var

ALPHA =

Alpha level for intervals

MU0 =

Parameter for t-test of the mean

HISTOGRAM / NORMAL

Bell curve overlay on histogram

QQPLOT

High-quality normal probability plot

INSET

Stats displayed inside the graph

💡

Normality check: Skewness ≈ 0 + Kurtosis ≈ 0 + NORMAL test p-value > 0.05 → data is approximately normal. Reference files: AUNI1.SAS, AUNI2.SAS, AUNI4.SAS

📌

PROC MEANS vs PROC UNIVARIATE: MEANS = quick summary table. UNIVARIATE = normality tests, distribution plots, outlier detection.

SAS Programming
Chapters 1 – 6

Getting Started with SAS

Getting Data into SAS

Reading, Writing & Importing Data

Preparing Data for Analysis

SAS Procedures

Evaluating Quantitative Data

SAS ProgrammingChapters 1 – 6

Getting Started with SAS

Getting Data into SAS

Reading, Writing & Importing Data

Preparing Data for Analysis

SAS Procedures

Evaluating Quantitative Data

SAS Programming
Chapters 1 – 6