How to Create Custom Formats in SAS with PROC FORMAT

You can use PROC FORMAT in SAS to define a precise mapping of data values into corresponding descriptive labels. This mechanism is crucial for condensing continuous data or categorizing discrete observations effectively.

Understanding the Basic PROC FORMAT Syntax

The fundamental usage of the PROC FORMAT procedure requires defining a format name using the VALUE statement and then specifying the input ranges or values and their corresponding output labels. The procedure must be terminated by a RUN statement.

The following example illustrates the basic syntax used to define a custom format named points_range, which categorizes numerical values into qualitative labels:

proc format;
    value points_range
        25-high='High'
        15-<25='Medium'
        other ='Low';
run;

This structure effectively establishes clear cut-off points for the data, translating quantitative scores into meaningful categories. Specifically, this definition creates the following categorical mapping logic:

• Values equal to 25 or greater (up to the highest value) will be displayed as the label ‘High‘.
• Values ranging from 15 up to, but not including, 25 will be shown as ‘Medium‘.
• All other numerical values that fall outside the defined ranges are captured by the OTHER keyword and will be shown as ‘Low‘.

Setting Up the Sample Dataset

To demonstrate the practical application of PROC FORMAT, we will utilize a small sample dataset containing player statistics. This data simulates scores (points) associated with different teams and positions.

The following SAS code uses the DATA step and DATALINES statement to create the temporary dataset named my_data, followed by the PROC PRINT step to display its contents:

/*create dataset*/
data my_data;
    input team $ position $ points;
    datalines;
A Guard 25
A Guard 20
A Guard 30
A Forward 25
A Forward 10
B Guard 10
B Guard 22
B Forward 30
B Forward 10
B Forward 10
B Forward 25
;
run;

/*view dataset*/
proc print data=my_data;

The resulting table, displayed below, shows the raw, unformatted point values for each observation:

Example 1: Applying Formats to Generate Descriptive Frequency Tables

One of the most common applications of PROC FORMAT is enhancing statistical reports, particularly when summarizing quantitative data. Before applying the format, let’s generate a simple frequency table for the points variable using PROC FREQ:

/*calculate frequency of values in points column*/
proc freq data = my_data;
    table points;
run;

As expected, the initial output displays the raw count and percentage of observations for each unique numerical value in the points column. While accurate, this raw output lacks descriptive context, especially if we are interested in performance bands rather than individual scores.

To make this output more insightful, we want to group these discrete points into performance categories. We will implement the previously defined categorization scheme:

• Values equal to 25 or greater will be displayed as ‘High‘ performance.
• Values between 15 (inclusive) and 25 (exclusive) will be shown as ‘Medium‘ performance.
• All other remaining values will be categorized as ‘Low‘ performance.

We combine the format definition step with the PROC FREQ step, using the FORMAT statement within PROC FREQ to apply the user-defined format points_range. to the points variable:

/*define formatting for points variable*/
proc format;
    value points_range
        25-high='High'
        15-<25='Medium'
        other ='Low';
run;

/*create frequency table for points variable, using formatting defined above*/
proc freq data = my_data;
    table points;
    format points points_range.;
run;

The resulting frequency table successfully utilizes the defined format labels, grouping the raw values of the points variable into aggregated categories. This transformation significantly enhances the clarity of the statistical summary.

Example 2: Creating a New Categorical Variable using the PUT Function

While the previous example demonstrated how to apply a format temporarily for reporting purposes (like in PROC FREQ), sometimes it is necessary to permanently store the formatted labels as a new categorical variable within the dataset.

This permanent transformation is achieved using the PUT function within a SAS DATA step. The PUT function reads the numerical value of the source variable (points) and applies the specified format (points_range.) to assign the corresponding character label to the new variable (point_range).

The following syntax defines the format and then executes the DATA step to create the new dataset, new_data:

/*define formatting for points variable*/
proc format; 
    value points_range
        25-high='High'
        15-<25='Medium' 
        other ='Low';
run;

/*create new dataset with points_range variable*/
data new_data;
    set my_data;
    point_range = put(points, points_range.);
run;

/*view dataset*/
proc print data=new_data;

The resulting output clearly shows the newly generated variable, point_range, appended to the original data. This new variable contains the textual labels (‘Low’, ‘Medium’, or ‘High’) corresponding directly to the numerical performance score in the points column.

Key Takeaways on Using PROC FORMAT

The examples provided illustrate the power and flexibility of PROC FORMAT in SAS programming. Whether you need to temporarily enhance the readability of reporting procedures like PROC FREQ or permanently categorize data using the PUT function to create a new dataset variable, user-defined formats are essential for effective data management and communication.

The ability to map complex numerical ranges to simple, descriptive text labels is vital for producing clear, actionable statistical output, especially when handling large volumes of raw data or generating reports intended for non-technical audiences.

Further Learning and Resources

For advanced options, such as defining picture formats for dates and times, or handling specific character formats, it is always recommended to consult the official documentation. You can find the complete and comprehensive documentation for PROC FORMAT at the official SAS website.

To deepen your expertise in data manipulation and statistical analysis within the SAS environment, consider exploring tutorials on related procedures such as PROC FREQ, PROC PRINT, or the DATA step syntax.

The following list provides references to tutorials that explain how to perform other common tasks in SAS:

• Tutorial 1: Using SAS to merge datasets.
• Tutorial 2: Advanced filtering techniques with WHERE statements.
• Tutorial 3: Understanding macro variables in SAS.