Sas How to Calculate Frequency Without Proc Freq

Why Use an Alternative to PROC FREQ?

While PROC FREQ is the standard SAS procedure for frequency analysis, there are situations where you might need to use alternative methods:

• When you need more control over the output format
• When working with very large datasets that PROC FREQ can't handle efficiently
• When you need to combine frequency calculations with other data processing steps
• When you want to create custom frequency tables that aren't available in PROC FREQ

The alternative methods we'll cover use SAS DATA steps and PROC SQL, which provide more flexibility while still being efficient.

Basic Frequency Calculation

The simplest frequency calculation counts how many times each value appears in a dataset. Here's how to do this using a DATA step:

DATA frequency_counts;
    SET your_dataset;
    BY variable_of_interest;
    RETAIN count;
    IF FIRST.variable_of_interest THEN count = 0;
    count + 1;
    IF LAST.variable_of_interest THEN OUTPUT;
RUN;

This code creates a new dataset with two variables: the unique values from your variable of interest and the count of how often each value appears.

Creating Frequency Tables

For more complex frequency tables, you can use PROC SQL to create cross-tabulations. Here's an example:

PROC SQL;
    CREATE TABLE frequency_table AS
    SELECT
        variable1,
        variable2,
        COUNT(*) AS frequency
    FROM your_dataset
    GROUP BY variable1, variable2
    ORDER BY variable1, variable2;
QUIT;

This creates a frequency table showing how often each combination of values from variable1 and variable2 appears in your dataset.

Example Calculation

Let's look at a practical example. Suppose you have a dataset of survey responses and want to count how many people selected each response option.

Sample Data

Calculating Frequencies

Using the DATA step method:

DATA response_frequencies;
    SET survey_data;
    BY response;
    RETAIN count;
    IF FIRST.response THEN count = 0;
    count + 1;
    IF LAST.response THEN OUTPUT;
RUN;

Resulting Frequency Table