Biostatistical Computing Table of Contents

BIS524 Chapter 1

Input and Infile: Building a SAS Data Set from Raw Data

Abbreviations

Introduction to SAS

Example job

Deconstructing the example

SAS steps

SAS Syntax: Generic rule

Rules for SAS names:

What is a SAS data set?

What is in the dataset? Data set information.

What is in the dataset? Variable attributes.

Function and syntax of DATA statement

Procedure (Proc) Step

Step boundaries

Reading raw data

Function of INPUT statement

Syntax of INPUT statement using LIST input

Where are the data?

Reading raw data separated by spaces

Job output of CP0101

Blanks as delimiters

Missing values

Output Window

Log Window

Moral: How to handle missing values when using LIST input

Reading data arranged in columns

Column oriented input

Input statement syntax for COLUMN input

Reading selected fields

Reading fields in different order

Reading column data that require informats

Syntax of FORMATTED input

Input buffer

Pointer control

SAS Informats

Some Common Informats

Dates

Dates

Dates (and other numbers)

Format statement

What goes on in the Data step?

The Compilation Phase

Compile time

Step 1

Step 2

Step 3

Step 4

Step 5

Step 6

Step 7

The results of this code after compile time

The Execution Phase

The results of this code during run time

Results at the end of the first execution pass through the data set

Results at the end of the last execution pass through the data set

Other examples using list input

Using informats with list input

Running the above code illustrates this common situation.

Inspecting the LOG

Notice what happens when SAS encounters invalid data.

Two ways to solve these problems

Syntax of the INFORMAT statement

Use of the INFORMAT statement

Use of the colon (:) modifier

Blanks in Data Fields

The following INPUT statement will work, however:

Values separated by something other than blanks

Reading multiple records (lines) per observation

Reading multiple records

Reading multiple records

Reading multiple fields

Output

Discussion

Abbreviating variable names

Using a variable list and Informat list

Which style of INPUT to use? List or Column or Formatted?

Advanced variable/informat lists

Reading a mixture of record types: conditional input

Example 10: Code

Understanding the INPUT statement.

Example 10: Discussion

Conditional Execution

The Full Example - 1

The Full Example - 2

The Full Example - 3

SAS Constants

An alternate coding.

One Million Points Off On Homework!

How do you choose between character or numeric variables?

Warning about character conversion?

How do you decide?

Deciding

A Fix?

A bottom line

Reading data from external files

Reading data from external files

More on FILENAME and INFILE

INFILE statement notes

Reading parts of raw data files

BIS524 Chapter 2

Data Recoding

Introduction

Example data

Example 1: Brute force recoding

The Assignment statement

Ask yourself (play compiler):

Does the code work?

Other forms of the IF-THEN statement

Style issues in use of the IF-THEN statement

Common mistakes in the use of the IF-THEN structure

Missing values

An alternate coding

Efficiency considerations

Example 2: The SELECT-WHEN structure

When to use IF-THEN/ELSE vs. SELECT-WHEN

From Example 10 in Chapter 1

The general SELECT/WHEN structure

Writing error messages to the SAS log

Recoding using Formats

Proc Format syntax

Proc Format syntax

This brings up some problems

Job Output

Follow thru each value of SCORE

Fixing the format

When first endpoint = previous last endpoint

Example 4 creating a new variable using formats

Using the PUT function

Ask yourself (act like you are the SAS compiler):

BIS524 Chapter 3

Reading and combining SAS data set

Review: Building SAS data sets

Syntax of a SAS file name

Building a permanent SAS data set

LIBNAME statement

The whole job

SAS vs. WINDOWS (WIN) file names

Using a permanent SAS data set

Using a permanent data set

What is in a SAS data set?

Proc Contents example

Good practice: Creating a permanent data set

Using a permanent SAS data set In the Data step

Bad (terrible!) practice

Combining SAS data sets

Concatenating data sets

What if the variables are in a different order?

What if the variable names don't match?

What if the variable types don't match?

What if the measures (variable values) don't match?

Subsetting SAS data sets

Subsetting observations

Subsetting, an alternate form

Choosing between subsetting IF and IF-THEN DELETE

Use of the Where statement

Differences between subsetting IF and WHERE

Allowable expressions in the WHERE statement

Using WHERE in Proc statements

Subsetting variables

Use of the Drop or Keep statements

Use of Drop= or Keep= data set options

Using both forms of Subsetting - Error?

Using both forms of Subsetting - Okay now?

Even Better

Merging data sets

One-to-one merges

Failure of the simple one-to-one Merge

Proc Sort syntax

Match merges

Possible problems: Data sets not sorted

Possible problems: Missing observations

Missing Observations

Possible problems: Identical variable names

Data set option: Rename syntax

Creating more than one data set with a single data step

Splitting one SAS data set into two data sets

Results of splitting the data set

OUTPUT Statement

Updating a SAS data set

Results of updating the master data set

UPDATE Statement

NAMED Input

BIS524 Chapter 4

Table Lookup Tools

What is a table lookup?

Table Lookup using Match Merging

Table Lookup using Formats, and Put & Input Functions

BIS524 Chapters 9-13

Table Lookup Tools

Chapter 9: Proc Print Syntax

Examples of using Proc Print

Chapter 10: Proc Means and Proc Univariate

Proc Means Syntax

Examples of using Proc Means

Proc Univariate Syntax

Examples of using Proc Univariate

Chapter 11: Proc Format and Proc Freq

Proc Format Syntax

Examples of using Proc Format

Proc Freq Syntax

Examples of using Proc Freq

Chapter 12: Proc Chart Syntax

Examples of using Proc Chart

Chapter 13: Proc Plot Syntax

Examples of using Proc Plot

BIS 524 Chapter 5

Precision of Numeric Variables

Let's add up 1/10 ten times

When is zero not zero?

Length of Numeric Variables

Floating Point Representation of Real Numbers

Floating Point Example - Storing an integer

Floating Point Example - Storing a Non-Integer

Numeric Precision - Storing an Integer

Numeric Precision - Take a Big Byte?

The LENGTH statement

Chapter 5: Functions

Example 1: Mathematical Transformations

Example 4: Statistical Functions

Example 6: Conversion of numeric values to character values

Example 5: Conversion of character values to numeric values

Example 3: Rounding and truncating

Rounding, reconsidered

Consider: Why round?

Example 2: Subsetting every nth Observation

Example 7: Moving Averages

Example 8: Operations on Character Values

Example 10: Reading Fields with a combination of Numeric and Character values

Example 14: Concatenating Character Strings

BIS 524 Chapter 6

SAS Dates

SAS Formats and Informats

Which Century Is It?

SAS Date from Month, Day, Year

SAS Date Constants

The Today Function

Counting the Days (or Months or Years)

Calculating a Person's Age

BIS 524 Chapter 7

SAS Arrays

Initializing a Group of Variables

Array Statement

Character Arrays

Array Containing All Numeric, All Character Variables, Or Temporary Variables

DO Structure

Do Group

Iterative DO Loop

Creating Several Observations from One Observation

BIS 524 Chapter 8

"Remembering" Values from Previous Observations

A Common Mistake

Remember The Values

RETAIN Statement - data step

First. and Last. Automatic Variables

SUM Statement and Implicit Retain

Example using RETAIN, DO, ARRAY and More

BIS 524 Chapter 14

Efficiency: Making Your Programs More Efficient

Large Data Sets - It is worth the effort to be efficient?

Compressed SAS Data Sets

Storing Data When Storing Summaries Will Suffice

Avoiding Unnecessary DATA Steps

Create a Data Set Using One Data Step Whenever Possible

Selecting Records with an IF statement: When do you use it?

Using a WHERE Statement to Select Records from a SAS Data Set

Dropping Unnecessary Variables When Building a SAS Data Set from Raw Data

Dropping Unnecessary Variables When Building a SAS Data Set from Raw Data

Review These Ideas (we've covered them before)

Logical Comparisons: Multiple OR Operators versus the IN Operator

Proc Datasets

Proc Datasets Syntax (partial - see documentation for full syntax)

Proc Datasets Example

Appending Data to a SAS Data

Appending Data with a Data Step

Appending Data with Proc Append

Proc Append with Matching Variables and Variable Types

Using Proc Datasets to Append Data

Copying a SAS Data Set from One Library to Another

Other Considerations