SAS Base Certification Mastery: Learning by examples

I regularly get questions from people who want to take SAS base certification about:

1. proper materials to prepare

2. how long it takes to prepare for the exam

3. whether prior experience with SAS is required

This course will offer a low stress approach to learning SAS and prepare for the exam.

It is built with learning by examples in mind and organized around the following themes:

     a. Step 1: How to read data into SAS (regardless of where it comes from)

     b. Step 2: How to explore data in SAS (review, prepare, understand)

     c. Step 3: How to transform data in SAS (new variables, merge, concatenate, subset, transpose…)

    d. Step 4: Understanding: Behind the scene in SAS and error handling (PROGRAM DATA VECTOR)

    e. Step 5: Building reports (we could have focused on statistics or visualizations)

    f. Step 6: Exporting reports and data outside SAS

Any code can be broken down into a combination of the following elements:

1. DATA STEP

    • Mostly used for data manipulation

2. PROC STEPS: ex: PROC PRINT, PROC FREQ, PROC MEANS…

   Use for:

     •Reporting

     •Graphics

     •Statistical analyses

3. PROC SQL

     •Alternative to DATA STEP and PROC STEP for relational databases. Not covered in this course

4. MACROS

    •This is SAS extension. It allows for building code that is more reusable

3. GLOBAL STATEMENTS and GLOBAL OPTIONS

    •Global statements and options are additional elements in SAS code that do not depend on any one of the     components mentioned above

DATA STEP is used to perform among other things the following tasks:

• reading SAS data sets (extension: .sas7bdat)

• reading external files including Raw data (txt, dat, csv) or proprietary sources (EXCEL, SPSS)

• subsetting data sets: that is trimming the data set size either by dropping some columns (keeping fewer columns) or filtering data (dropping some rows)

• Combining data sets: Merging or concatenating

• Transforming data: adding new variables, changing values, recoding variables

• Exporting data to external sources

The goal in this Use case is to:

• create a basic research data set by combining two other data sets: one demographics information (demo1) and another  containing income information (demo2)

• allows to see at a basic level how SAS components can be used together to answer a real life data management question

Type of data manipulations that will be covered include:

• cleaning data

• formatting data

• labelling data

• merging data

• creating reports/statistics

This lecture walks you through the basic steps of using DATA STEP in SAS code

This lecture illustrates how DATA STEP can be used to answer real life question

PROC STEPS or PROCEDURES are used for:

• Building reports: proc report, proc tabulate; proc print

• Creating graphics: PROC SGPLOT, PROC GPLOT

• Performing statistical analyses: PROC REG, PROC GLM…

• Reviewing data quality: PROC PRINT, PROC FREQ, PROC MEANS

• And many other tasks

• PROC STEPS are the most common in SAS code but are the least difficult to learn

This lecture illustrates the use of PROC STEPS in SAS code

This lecture shows examples of how to use PROC STEPS to answer real life questions

Macros in SAS are used to expand SAS capabilities, more specifically, they are used for:

• Handling repetitive code

• Inserting complex logic in SAS code and make it more generic

There are 4 main ways of creating macros (that is macro variables or macros):

• %let statement

• CALL SYMPUT

• PROC SQL, SELECT INTO:

• %MACRO……..%MEND;

Only %LET STATEMENT is covered in this course

This lecture illustrate how and why to implement %LET STATEMENT in SAS

This lecture shows examples of how %LET STATEMENT can be used to answer real life questions

Global statements and global options are additional elements in SAS not classifiable elsewhere

Examples of global statements include:

• LIBNAME: used create libraries (for accessing SAS or some other data sets)

• Title or Footnote: used to provide descriptive context to reports

• ODS and ODS close: used to send outputs to destination reports: PDF, EXCEL, WORD...

Examples of global options:

• OBS=...: Used to limit records to print from a DATA STEP or PROC STEP

• DATE/NODATE: used to decide whether to print the current date on a report or not

SAS Components: Putting all together

This lecture confirms that whatever your data management question is, the SAS code you need to answer it is always a combination of the components we covered in this section

The following will be covered:

• STEPS: the core components in SAS

• STATEMENTS and options: the elements that make up steps and other components

• SAS Naming rules: data set, variable, library

• Commenting code in SAS: one liner vs multiple lines comments

A step is either a DATA STEP or PROC STEP

(1) A DATA step starts with the keyword DATA and usually ends with RUN

(2) A PROC STEP starts with the keyword PROC and usually ends with RUN

Examples of what statements do:

• drop or keep variables: DROP/KEEP

• filter records: WHERE statement or IF  statement

• create new variables

Examples of what options do:

• Drop or keep variables: (DROP=) or (KEEP=)

• OPTION OBS=n

• Option NODATE/DATE

How to:

• name data sets or variables

• name libraries

• use default naming rules vs extended rules.

How to comment code in SAS:

• Use one liner comment: asterisk followed by comment followed by semicolon

• Use multiple liners comment: slash followed by asterisk followed by comments followed by asterisk followed by slash

Why to comment code in SAS:

• to  provide context about what the code is doing

• to help with understanding of the code by other users

• to ensure that SAS does not attempt to interpret comments

3 main ways of reading data into SAS:

• Using library/Engine option

• Using INPUT statement (used for non SAS data sets)

• Using PROC IMPORT

More about:

• Library/engine option

• INPUT STATEMENT option

• PROC IMPORT option

Ways for reading data into SAS:

• using LIBNAME with engine V9 for SAS DATA SETS

• Using LIBNAME with XLSX engine for excel files

• Using PROC IMPORT for various files types: excel, txt, csv, dat...

• Using INFILE/INPUT or INPUT/INFILE

Reading SAS DATA SETS (extension: .sas7bdat) using LIBRARIES:

• example: LIBNAME ABC  V9 "path to file";

• ABC is the library name

• path to file: is the folder path to where the SAS data set is located

• V9: the engine is optional

Reading data using excel libraries:

• example: libname ABC xlsx "path to excel file\excelfile.xlsx";

• ABC is the library name

• path to excel file is the folder path to the excel file

• excelfile.xlsx is the target file

• Each sheet in the file will represent a new SAS DATA SET

• XLSX is the engine. It is required

Reading files of types:

• CSV

• EXCEL

• TEXT, DAT

Illustration of how to use PROC IMPORT to read data into SAS

INPUT STATEMENT can be used to read any non SAS data into SAS

• It is the tool of last resolve when other options fail

• it uses INFILE to access data when the data is saved in external files or

• DATALINES OR CARDS when data is captured in the code editor

• it does depend on user declaring variables and their attributes (type, length, informat...).

• Various input styles are used depending on how the data was formatted or prepared in the source

  • LIST INPUT: data columns are separated by a common delimiter

  • COLUMN INPUT: data is captured in well pre-defined  columns easily separable

  • FORMAT INPUT: customed data captured that requires user to control the pointer to read different portions of data

The following procedures are essential for data exploration:

• PROC CONTENTS: for reviewing variables and their attributes

• PROC PRINT: for printing any portion of data

• PROC FREQ: for creating a summary count of data per variable

• PROC MEANS: for creating summary statistics on continuous variables

PROC CONTENTS is the best tool used to learn about the variables in data

if variables need description:

• use LABEL STATEMENT

if values in a variable needs description

• use FORMAT statement

Although SAS has a lot of formats integrated (ex: Date formats), once in a while a user-defined format is needed:

• example: if GENDER is captured as 1 and 2 (where 1 corresponds to female and 2 to male), a user-defined format can be used to capture this as Female and Male in reports, statistics and graphics

VALUE STATEMENT is one way to do this

• Value statement option is more appropriate when the list of values to format is small and not prone to changes

• a PROC FORMAT is used.

A second way to create user-defined formats is CNTLIN=option

• while VALUE STATEMENT OPTION requires that user manually matches the values to format to the labels to use as descriptions,

• CNTLIN=option takes all of the information needed (values, labels, format name) from a data set

• CNTLIN=option is more appropriate when the list of values to format is large or changes more often

Some additional options under PROC CONTENTS:

• Printing variables in the same order as in the data set: VARNUM

• Printing only the variables list: SHORT

• Outputting the variables list to a new data set: OUT=

PROC PRINT is used as an alternative to opening the data set

• VAR statement is used to control which variables print

• (FIRSTOBS=)/(OBS=) and WHERE are used to control which observations print

This lecture will explore the essential options under PROC PRINT

PROC FREQ allows for learning about participants in the data, one variable at a time such as:

• how many participants carry a given characteristic of the sample of a variable

• what percentage of the sample carries such a characteristics on the given variable

• the ability to cross-tabulate two or more variables

This lecture offers the opportunity to explore essential options under PROC FREQ procedure

PROC MEANS is designed to:

• build statistics from numeric variables; statistics that include the mean, median, standard deviation...

• breakdown such statistics by the values of a categorical variable

• PROC MEANS should not be run on numeric variables that do not a numeric meaning

This lecture provides an overview of the use of PROC MEANS

Combining data can consist of :

• concatenating or vertical combining: stacking 2 or more data sets on top of one another

• Merging or horizontal combining: more often matching records by a common matching variable or ID

This lecture provides a high level description of DATA CONCATENATING and the requirements for successful combining

This lecture offers a description of what MERGING is about what requirements must be met to succeed

While combining data consisted of bringing different data sources into one, data subsetting is designed to reduced data size.

It can be done by reducing the columns in the data or by reducing the rows

Subsetting by columns is done by either dropping columns not needed or keep columns of interest

Subsetting data consists of reducing the number of rows:

• by using OBS= or its variants

• WHERE or IF conditions in DATA STEPs

• using WHERE in procedures

Examples of data transformation include:

a. Create variables

• using assignment statement

• Using length statement

• using functions

• using conditional statements

b. Recoding variables

• using conditional statements

• using INPUT and PUT functions along with a format

c. Transpose data

• using PROC TRANSPOSE

• DATA STEP  or + arrays

d. Formatting data

The use of variable assignment statement and/or length statement for creating new variables

The use of conditional statements:

• IF THEN

• IF THEN DO

How to create user-defined formats:

• The use of the VALUE STATEMENT

• The use of the CNTLIN=OPTION

  
  

Data accumulation:

• by using SUM STATEMENT or SUM function and RETAIN

• by using FIRST.VARIABLE and/or LAST.VARIABLE to break down by categories

Transposing data:

• long to wide format or vice versa

• Using PROC TRANSPOSE

• an alternative is to use DATA STEP and arrays

Converting data from numeric to character or vice versa using:

• INPUT or PUT functions and

• A FORMAT