*Copyright @ www.statschoice.com;

/********************************************************************************

What is the word meaning of subset?

********************************************************************************/

/********************************************************************************

a part of a larger group of related things.

********************************************************************************/

/********************************************************************************

Where do we need to subset data?

********************************************************************************/

/********************************************************************************

Subset the records of cars manufactured in a particular continent
Select only the male student records from the class dataset
Identify and list the subjects of a clinical trial whose age is grater then 60

********************************************************************************/

/********************************************************************************

Create sample input datasets

********************************************************************************/

data class;
     set sashelp.class;
  run;
   
  data cars;
     set sashelp.cars;
  run;

/********************************************************************************

Subset the records using 'where' statement

********************************************************************************/

/********************************************************************************

To subset the observations, we need to specify the record selection condition on the where statement

********************************************************************************/

/********************************************************************************

Create a dataset named males by subsetting the male student records from class dataset

********************************************************************************/

/********************************************************************************

Notice in the log about the number of observations read from the input dataset
With where statement, only those records which meet the specified subset condition will be read into PDV from the input dataset

********************************************************************************/

data males;
     set class;
     where sex="M";
  run;

/********************************************************************************

Create a dataset named females_preteen by subsetting female student records whose age is less than 13 from class dataset

********************************************************************************/

/********************************************************************************

We need to filter the records based on the values present in two different variables: SEX and AGE
We can combine the filter components with an 'and' operator

********************************************************************************/

data females_preteen;
     set class;
     where sex="F" and age lt 13;
  run;

/********************************************************************************

Create a dataset named cars_mileage by subsetting the cars with mpg_city > 24 or mpg_highway >30 from cars dataset

********************************************************************************/

/********************************************************************************

We ned to filter the records based on teh values present in two different variables: MPG_CITY and MPG_HIGHWAY
As we need to filter the records which meet at least one filter condition, we can combine the filter combponents with an 'or' operator

********************************************************************************/

data cars_mileage;
     set cars;
     where mpg_city gt 24 or mpg_highway gt 30;
  run;

/********************************************************************************

Subsetting the records with subsetting if statement

********************************************************************************/

/********************************************************************************

When a subsetting if statement is used, SAS continues processing only those observations that meet the condition of the specified expression.
When the specified expression gets evaluated to false, SAS returns the control back to data statement and reads next observation into PDV
AS SAS writes an observation to the output dataset only when a step boundary is encountered, the observations which do not meet the filter expression never reaches the step boundary in the execution, hence those observations will never be written to the output dataset

********************************************************************************/

/********************************************************************************

Create a dataset named females by subsetting female records using subsetting if statement

********************************************************************************/

data females;
     set class;
     if sex="F";
  run;

/********************************************************************************

Create a dataset named select_sedans by subsetting Sedan cars with a highway mileage greater than 35 using subsetting if statement

********************************************************************************/

data select_sedans;
     set cars;
     if type="Sedan" and mpg_highway gt 35;
  run;

/********************************************************************************

Create a dataset named cheap_mileage by subsetting the cars with msrp < 15000 dollars or city mileage greater than 40 using subsetting if statement

********************************************************************************/

data cheap_mileage;
     set cars;
     if msrp<15000 or mpg_city gt 40;
  run;

/********************************************************************************

Subsetting the records using if ... then output

********************************************************************************/

/********************************************************************************

Output statement in SAS is used for instructing the data step to write the contents of PDV to the output dataset
In this case, we are making use of conditional processing (if then syntax) along with output to subset the observations
With the syntax, 'if (condition is true) then output;' we are instructing SAS to write only those observations to the output dataset which meet a specific filtering condition

********************************************************************************/

/********************************************************************************

Create a dataset named tall by subsetting the students with height greater than 60 using if ... then output

********************************************************************************/

data tall;
     set class;
     if height gt 60 then output;
  run;

/********************************************************************************

Subsetting the records using if ... then delete

********************************************************************************/

/********************************************************************************

Delete statement in SAS is used in instructing the data step to delete the contents of PDV and fetch next observation for processing, in turn the current observation is not written to the output dataset
With the syntax, 'if (condition is true) then delete;' we are instructing SAS to stop writing the specified observations to the output dataset which meet a specific filtering condition

********************************************************************************/

/********************************************************************************

Create a dataset named short by subsetting the students with height less than 60 using if ... then delete

********************************************************************************/

/********************************************************************************

Notice that as we need student records with height less than 60, we are deleting the records of students whose height is greater than or equal to 60

********************************************************************************/

data short;
     set class;
     if height ge 60 then delete;
  run;