To Lecture Notes

CSC 433 -- Mar 28, 2016

General Comments about CSC 433

• Prior knowledge of SAS or R is not assumed, although we will cover these languages fairly quickly.
   
• Prior knowledge of advanced statistics is not assumed. You need not have taken CSC 423.
   
• The prerequisite for CSC 433 is some programming experience, for example a first programming course like CSC 401. We will also use concepts from exploratory data analysis, for example, IT 403. Concepts such as means, standard deviations, percentiles, and boxplots will be used.
   
• CSC 433 will cover both SAS and R: SAS in the first half; R in the second half.

Review Questions

1. Discuss the similarities and differences between SAS data and proc steps.
    
2. Find the mistakes in this SAS code:
   

   // Create dataset.
   data quiz-scores;
      input name score;
      datalines;
   William 95
   Jenna   92
   Scott   89
   Carla   94
   
   proc means / mean;
   

3. Using the input file kids.txt, write a SAS program that does the following:
    
   
   1. finds the total number of kids,
       
   2. finds the minimum and maximum ages of the kids,
       
   3. finds the minimum and maximum ages as well as the total number of the girls and boys separately,
       
   4. finds the first girl and boy names in alphabetical order.
   
   
4. For what are SAS labels used?

Obtaining Details about a Dataset

• Use proc contents to obtain details about a SAS dataset.
   
• The Explorer Window in the SAS IDE can provides dataset details similar to proc contents.
   
• Problem: try out proc contents on the kids dataset. What information do you obtain?

Getting Started with SAS

• SAS executes line by line and observation by observation.
   
• The lines of a SAS data step are executed at least once.
   
• An implicit while loop is present in a datastep that is reading from an input data file.
   
• The while loop repeats until the input data file runs out of observations.
   
• For many statistical procs, ODS (Output Delivery Service) is produced automatically.
   
• Try running this script:
  

  data test;
      input x y @@;
      datalines;
  1 1  2 3  3 2  4 4
  ;
  
  proc reg;
      model y = x;
      
  run;
  quit;
  

More SAS Questions

1. Create a SAS dataset without using an input statement. How can you create more than one observation in this way?
   Ans: To create a dataset with one observation:
   

   data kids;
      name = "Alice";
      gender = "F";
      age = 11;
   

   To create a dataset with more than one observation:
   

   data kids;
      name = "Alice";
      gender = "F";
      age = 11;
      output;
      name = "Scott";
      gender = "M";
      age = 12;
      output;
   

   The default behavior is to send the data the data step to the dataset automatically each time the end of the data step is reached. If you want to send an observation to the dataset in the middle of the the data step, use the output statement. However, if you use any output statements in a data step, you must use them every time you want to send an observation to the dataset, even at the end of the data step.
    
2. Here are three data files, quiz1.txt, quiz2.txt, and quiz3.txt each containing 8 quizzes, a midterm and a final exam for each student. In quiz1.txt, one input line is used to generate one observation.  In quiz2.txt, each line is used to generate three observations. Finally, in quiz3.txt, three input lines are used to form each observation.  Use the specifications in these tables to compute the course score and final grade for each of the input files.
    

   Score Proportion of Course Score Quiz Average 40% Midterm 30% Final 30%

   Course Score Course Grade 90 - 100 A 80 - 90 B 70 - 80 C 60 - 70 D 0 - 60 F

   
   Use @@ at the end of an input line to hold the line until all the data on that line is read.
    
   Use @ at the end of an input line to hold the line until the end of the data step.
    
   The SAS symbol for missing value is a period ( . ).
    
   Answers: See the Quiz Examples (Quiz1, Quiz2, and Quiz3).
   

Specifying the Read Position in an Input File

Pointer	Meaning
m-n	Read data from column n to column n.
n	Read one character from column n.
@n	Move pointer to column n for subsequent reading.
+n	Move pointer n characters forward for subsequent reading.
/	Move to next line of input.
#n	Move to line n of input.
@'chars'	Move to past 'chars' character string.



• Note: if @n or +n is used to position the column pointer at column m, the next read will occur at column m+1.
   
• See the Pointer and WebLog Examples.

SAS Informats

• A SAS informat is a string of characters that represents instructions for reading data into a variable.
   
• A table of common SAS informats.
   
• See the SAS Language Manual for a more complete discussion of informats.
   
• Use this Table of Extended ASCII Characters to understand the following Informat Examples.
   
• Look at the Informat Examples: States, USPopulation, HexChars
   
• Some comments about informats that were mentioned in the lecture:
   
  
  1. SAS only has two datatypes: character and numeric.  All other "datatypes" are interpretations of these two types.
      
  2. To represent the boolean values false and true, use the numeric values 0 and 1, respectively.
      
  3. The SAS zero date is January 1, 1960.
      
  4. The colon prefix ( : ) modifies an informat to stop when a blank is encountered.  For example:
      
     
     $CHAR10. is a character informat that reads a field of with 10, retaining leading, embedded, and trailing blanks.
      
     :$CHAR10. is a character informat that reads a field of with 10, but stops reading when a blank is encountered. The pointer is positioned to the next character after the first blank encountered.
  
  
• We will discuss the Sound Example next week.

SAS Formats

• A SAS format is a string of characters that is used for writing output in the correct format.
   
• A table of common SAS formats.
   
• See this paper by Andrew Carp for more details on SAS formats.

Writing to a File

• Use the file and put statements to write to an output file on disc.
   
• A data step with the name _null_ is traditionally used when using the file and put statements so that no actual dataset is created.
   
• See the PutToFile and Sales Examples.
   
• The Sales Example uses first and last variables to facilitate printing regional totals. It uses first and last variables that are automatically generated with the set statement with a by variable. See the FirstLast Example.

Project 1

• Look at the Project 1 Description.