How to display values in percent format in SAS?

Understanding Percentage Formatting in SAS

Numerical data in SAS is inherently stored without predefined formatting instructions. When you load data that represents proportions or fractions—such as exam scores, market share, or response rates—these values exist as floating-point numbers between 0 and 1. To convey the meaning of these numbers effectively to non-technical stakeholders, it is essential to apply a display format that multiplies the stored value by 100 and appends the ‘%’ symbol.

The PERCENT format option in SAS is specifically designed to handle this conversion automatically. It is a powerful tool because it controls only the appearance of the data during viewing or printing; it does not alter the underlying numerical value stored in the Dataset. This distinction is crucial: calculations can be performed on the raw proportional values, while reports display the formatted percentages.

The flexibility of the PERCENTw.d format allows users to tailor the output to specific reporting requirements. Whether the need is for high precision (many decimal places) for scientific reports or rounded integers for executive summaries, the format parameters offer complete control. Understanding how the width (w) and decimal places (d) interact is the key to generating clean and professional output in any SAS application, particularly within reporting procedures like PROC PRINT.

The Anatomy of the Percentw.d Format

The Percentw.d Format is defined by two crucial components: w and d. The w value, or the width parameter, dictates the total number of columns reserved for displaying the formatted percentage. This total width must accommodate the numerical digits, the decimal separator, the percentage sign (%), and, if applicable, the negative sign. If the specified width is too small to display the full value, SAS will often display asterisks (*) to indicate truncation, which is why prudent width selection is important.

Conversely, the d value, or the decimal parameter, precisely controls the number of digits that appear to the right of the decimal point. This value determines the level of precision shown to the reader. When SAS applies the format, it automatically multiplies the internal fractional value by 100. It then ensures that the resulting percentage is rounded to the d number of decimal places specified, following standard rounding rules (rounding up if the next digit is 5 or greater).

For instance, consider the value 0.945. If formatted using PERCENT7.1, the output would be 94.5%. The width (7) accommodates the three digits, the decimal, the one decimal place, and the percent sign (e.g., 94.5%). If formatted using PERCENT7.0, the output would be 95%. In this case, d=0 triggers rounding the percentage (94.5) to the nearest integer (95), and the total width is still sufficient (e.g., 95%).

Choosing the correct combination of w and d is essential for both accuracy and aesthetics. If w is significantly larger than necessary, the resulting column may contain excessive white space, impacting report compactness. If d is set too high, it might introduce unnecessary precision that distracts the reader. Therefore, selecting a format like PERCENT10.2 often represents a good balance, providing ample space (10 characters) and standard two-decimal precision.

Setting Up the Sample Dataset in SAS

To demonstrate the practical application of percentage formatting, we will first create a simple Dataset containing hypothetical exam scores. These scores are stored as fractional values, ranging from 0 to 1, representing the proportion of correct answers for several students in a class. This setup mirrors many real-world scenarios where raw scores must be converted for presentation.

The following code block utilizes the DATA step to generate a dataset named my_data. We use the INPUT statement to define two variables: student (character type, denoted by $) and exam_score (numeric type). The subsequent DATALINES statement holds the raw, unformatted data which will be processed by the SAS system.

We then conclude the data creation process with the RUN statement and immediately follow it with a simple PROC PRINT step. This initial print procedure serves to visualize the raw data as it is stored internally, confirming that the scores are indeed fractions before any formatting is applied. This baseline view is essential for verifying the impact of the subsequent formatting steps.

/*create dataset*/
data my_data;
    input student $ exam_score;
    datalines;
Andy 0.945
Bob 0.78
Chad 0.865
Derrick 0.77
Eric 0.75
Frank 0.64
George 0.895
Henry 0.98
Isaac 0.68
John 0.84
;
run;

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

Applying Percent Formatting with Decimal Precision

Once the baseline data structure is established, the next logical step is to apply the desired percentage Format to the exam_score column. The goal here is to convert the raw proportional values (e.g., 0.945) into visually clear percentages (e.g., 94.5%). For initial reporting, maintaining at least one decimal place is often preferred to ensure a higher degree of accuracy is conveyed without overcomplicating the display.

We accomplish this formatting within the PROC PRINT step using the FORMAT statement. The syntax is straightforward: specify the variable name followed by the desired format structure. In this example, we use percent10.1. This statement applies the percentage conversion and ensures that the output will reserve 10 characters for the field width and display exactly one digit after the decimal point.

The chosen format, percent10.1, is highly illustrative of the format’s power. The 10 provides generous space for scores ranging from 0% to 100% (e.g., 100.0% uses 6 characters), ensuring no truncation occurs, even with small fluctuations. The 1 digit of precision strikes a balance between detailed reporting and clear presentation. Observe how the original fractional values are transformed in the output generated by the code below.

/*view dataset and display exam scores in percent format*/
proc print data=my_data;
    format exam_score percent10.1;
run;

Controlling Output Width and Decimals

When reviewing the results of the previous step, every value in the exam_score column is now correctly displayed in a percent format. This transformation highlights the two roles played by the parameters in percent10.1. The overall width, defined by 10, guarantees that the entire formatted value—including the implicit multiplication by 100 and the addition of the percent symbol—fits within the reserved column space in the PROC PRINT output.

More specifically, the parameter 1 dictates the precise number of decimal places shown. If the underlying data (e.g., 0.945) results in a percentage (94.5%) that naturally requires fewer digits than the specified precision, SAS will still ensure that the exact number of decimal places is displayed, padding with zeros if necessary (though not required in this example). Conversely, if the unformatted percentage has more decimal places (e.g., 94.53%), SAS rounds the value to meet the d specification (94.5%).

Choosing an appropriate width (w) is primarily about preventing truncation errors, especially when dealing with extreme values (e.g., percentages over 100% or very large negative percentages). While 10 characters might seem large for scores between 0% and 100%, it provides a safety buffer. Furthermore, using the same width consistently across multiple reports aids in visual alignment and professional presentation, ensuring that all columns maintain a standard visual footprint regardless of the exact numerical content.

Displaying Percentages Without Decimal Places

In many summarized reports or dashboards, excessive precision can clutter the visual presentation. When the requirement is to show percentages rounded to the nearest whole number, we must modify the d parameter within the Percentw.d Format. To suppress the display of any fractional components, the decimal parameter is simply omitted, or explicitly set to zero (d=0).

If we use percent10. (omitting the decimal parameter entirely), SAS defaults to displaying zero decimal places. This command triggers two actions: first, the proportional score is converted to a percentage; second, this percentage is rounded to the nearest integer. It is vital to note that this is rounding, not simple truncation. For example, 94.5% is rounded up to 95%, while 77.0% remains 77%.

This rounding behavior is critical for accurate reporting. While the visual output is simplified, the data integrity is maintained through standard arithmetic rounding rules applied during the formatting process. The width (10) remains adequate for the display, allowing the resulting integer percentage value (e.g., 95%) and the percent symbol to fit comfortably within the column boundaries. The following code demonstrates this application and the resulting output.

/*view dataset and display exam scores in percent format without decimal places*/
proc print data=my_data;
    format exam_score percent10.;
run;

Upon reviewing the final output table, it is clear that each numerical value in the exam_score column has been rounded to the nearest integer. For instance, Andy’s score, originally 94.5%, is now displayed as 95%. This process confirms that SAS applies standard rounding when the decimal parameter is suppressed or set to zero, ensuring that fractional parts are handled appropriately for high-level reporting.

Key Considerations for SAS Format Usage

While the PERCENTw.d format is ideal for immediate reporting using procedures like PROC PRINT, advanced users should understand where and how formats are permanently stored and applied. The use of the FORMAT statement within a reporting procedure (as shown in the examples above) applies the format only for the duration of that procedure run. The underlying Dataset remains unchanged, and the variable retains its raw numerical values.

For scenarios where a format must be permanently associated with a variable, the FORMAT statement should be included within the DATA step when the dataset is created or modified. Alternatively, if custom formatting rules or multiple conditional formats are required, users can leverage PROC FORMAT to define specialized formats that can then be applied universally across different procedures and datasets.

The PERCENT format is one of many standard numeric formats available in SAS. Others include COMMAw.d for currency and large numbers, DATEw. for date interpretation, and DOLLARw.d. Mastery of these formatting options is a fundamental skill in SAS programming, ensuring that data presentation is accurate, professional, and tailored to the intended audience.