Proc Report Secreates Essay Example

Proc Report Secreates Essay PharmaSUG 2012 Paper TF20-SAS PROC REPORT Unwrapped: Exploring the Secrets behind One of the Most Popular Procedures in Base SAS ® Software Allison McMahill Booth, SAS Institute Inc. , Cary, NC, USA ABSTRACT Have you ever wondered why a numeric variable is referenced in different forms within a COMPUTE block? Do you know the difference between a DATA step variable and a variable that is listed in the COLUMN statement? Then, this paper is for you! Welcome to PROC REPORT Unwrapped. We are looking at PROC REPORT and uncovering some of the behind-the-scenes details about this classic procedure. We will explore the components associated with PROC REPORT and discover ways to move column headings and change default attributes with styles and CALL DEFINE statements. We will also dig deep into example code and explore the new ability to use multilabel formatting for creating subgroup combinations. So for anyone who has ever written PROC REPORT code, stay tuned. Its PROC REPORT Unwrapped! INTRODUCTION Which popular SAS procedure has features of the PRINT, MEANS, and TABULATE procedures and features of the DATA step in a single report-writing tool? It enables you to create a variety of reports including a detail report, which contains a row of data for every input data set observation, or a summary report, which consolidates data so that each row represents multiple input data set observations. Here is another hint: this same procedure provides the ability to create both default and customized summaries, add text and statistics, and create columns of data that do not exist in the input data set. If you guessed PROC REPORT, you are correct! We will write a custom essay sample on Proc Report Secreates specifically for you for only $16.38 $13.9/page Order now We will write a custom essay sample on Proc Report Secreates specifically for you FOR ONLY $16.38 $13.9/page Hire Writer We will write a custom essay sample on Proc Report Secreates specifically for you FOR ONLY $16.38 $13.9/page Hire Writer For anyone who has written PROC REPORT code and has wondered what is going on behind the scenes, this is the paper for you. This paper explores some of the behind-the-scenes secrets of PROC REPORT. We will dig deep into example code as we begin to uncover some of the details of this classic report-writing procedure. As a bonus, you will discover some facts about the REPORT procedure that you might not have known. By the way, the code output in this paper is based on the SAS ® 9. 3 default output destination of HTML. Although most of the paper ontent can also be applied to the LISTING destination, the code that is shown in this paper is intended to be used in an Output Delivery System (ODS) destination, unless otherwise indicated. With that being said†¦are you ready to explore? Welcome to PROC REPORT Unwrapped! EXPLORING THE SECRETS (HOW IT’S MADE) PROC REPORT first began life as a procedure many years ago in SAS ® 6. Since then, it has been gaining popularity as the t ool of choice for report writing. Even with such popularity, there are still aspects of the REPORT procedure that can be further explored. In this segment, we will unwrap and explore some of the secrets behind this most popular procedure with a focus on the following components: ? referencing a numeric variable in a COMPUTE block ? exploring the difference between an input data set variable and a DATA step variable ? discovering ways to move column headings ? changing default attributes with styles ? using the CALL DEFINE statement ? exploring the new ability in SAS 9. 3 to use multilabel formatting for creating subgroup combinations Let’s start exploring the secrets! REFERENCING A NUMERIC VARIABLE IN A COMPUTE BLOCK All numeric variables are referenced the same way, right? Well, that depends on how the numeric variable is defined in the PROC REPORT DEFINE statement. Before we can explore more about the how a numeric variable is defined, we first need to understand some PROC REPORT basics. Then we will explore the many ways a numeric variable 1 PROC REPORT Unwrapped: Exploring the Secrets behind One of the Most Popular Procedures in Base SAS ® Software, continued can be defined in the DEFINE statement and how that definition determines the manner in which the variable is referenced in a COMPUTE block. In the PROC REPORT statement, the input data set is listed using the option DATA= . If the DATA= option is not specified, PROC REPORT will use the last data set that was created in the current SAS session. The input data set contains variables and observations. The variables are categorized as either character or numeric— that is it, character or numeric. PROC REPORT does not use all of the variables from the input data set. Only the input data set variables that are listed in the COLUMN statement or in the BY statement are used. All of the report items, including the variables from the input data set that are listed in the COLUMN statement can be used in a COMPUTE block. Each report item in the COLUMN statement has an associated DEFINE statement. If a DEFINE statement for the report item is not supplied, PROC REPORT will create a default DEFINE statement behind the scenes. If a COLUMN statement is not specified, PROC REPORT will create a COLUMN statement behind the scenes. The COLUMN statement will contain only the variables from the input data set in the order of the data set. DEFINE statements can be supplied without a supplied COLUMN statement. The minimum statements that are needed to run PROC REPORT are a PROC REPORT statement with an input data set and a RUN statement. Behind the scenes, PROC REPORT will create all the necessary minimum default statements. To see the default statements, add the LIST option in the PROC REPORT statement. The LIST option will produce the basic code, including all of the DEFINE statements, in the SAS log. The NOWD option enables the report to run in the non-windowing mode. Here is an example of PROC REPORT code with the LIST option: proc report data=sashelp. class nowd list; run; The SAS log is shown in Output 1. Output 1. SAS Log Output By default, the DEFINE statement for a numeric input data set variable that is listed in the COLUMN statement will be associated with the SUM statistic. An alias for the SUM statistic is ANALYSIS. The SUM statistic is the most common statistic that is used in PROC REPORT code. The SUM statistic can be replaced with any valid PROC REPORT statistic such as MIN or MEAN. At BREAK and RBREAK rows, the numeric input data set variable with an associated statistic will consolidate automatically based on the associated statistic. When a numeric input data set variable with an associated statistic is referenced in a COMPUTE block, the form of the variable-name. statistic is used. In a COMPUTE block, if a numeric input data set variable name is used without the corresponding statistic (which is the statistic listed in the DEFINE statement), a note might be written to the SAS log. The following code will produce a note in the SAS log: roc report nowd data=sashelp. class; col age height weight total; define age / group; define heightweight/ mean; define total / computed; compute total; total=height. mean/weight; endcomp; run; 2 PROC REPORT Unwrapped: Exploring the Secrets behind One of the Most Popular Procedures in Base SAS ® Software, continued In the preceding code, the DEFINE statement for the WEIGHT variable lists MEAN as the statistic. The calculation in the COMPUTE TOTAL block for the TOTAL COMPUTED variable shows the WEIGHT variable without the statistic of MEAN. PROC REPORT requires this statistic and does not recognize the WEIGHT variable. A note, such as the following, is produced in the SAS log: NOTE: Variable weight is uninitialized. PROC REPORT allows duplication of report items in the COLUMN statement. This duplicated report item becomes an alias. When an alias of the numeric input data set variable is referenced in a COMPUTE block, the alias name is used without the associated statistic. Behind the scenes, any duplication of the same variable or statistic in the COLUMN statement will be associated with an alias name. If an alias name is not specified, PROC REPORT will create one. To see the assigned alias name, add the LIST option to the PROC REPORT statement and review the SAS log for the code. Using the preceding code in this section, the HEIGHT variable is duplicated in the COLUMN statement as follows: col age height height weight total; The resulting SAS log is shown in Output 2. Output 2. SAS Log Output Showing an Alias Name of _A1 Assigned behind the Scenes When the numeric input data set variable with an associated statistic is associated with an across variable, the column number, in the form of Cn_, is used in a COMPUTE block. In the form of _Cn_, n is the column number. The position of the columns shown in the output report is based on the left-to-right placement of the report-items in the COLUMN statement. For example, if a numeric variable with an associated statistic is placed as the first column under the ACROSS grouping but it is the second column in the output report, _C2_ is the correct value to use in a COMPUTE block. Behind the scenes, all columns are considered to have a column number even if the column is not printed in the final output report. Here is an example COLUMN statement: col sex age, (weight height); In this column statement, the first value of the WEIGHT variable is in the second column in the report. AGE is an across variable and is not counted as a column. The first column of the WEIGHT variable is associated with the first value of AGE and is referenced in a COMPUTE block as _C2_. The next column of the WEIGHT variable that is associated with the second value of AGE is referenced in a COMPUTE block as _C4_. Each unique value of the across variable becomes a header. Under each ACROSS header are the columns of variables that are associated with each unique across variable value. Each variable associated with an across variable becomes a column under the unique variable value. The number of unique values of an across variable controls the number of columns that are created for a variable associated with the across variable from the COLUMN statement. Behind the scenes, PROC REPORT has to know the specific column placement of a variable that is referenced in a COMPUTE block. The _Cn_ is used instead of the variable-name. statistic, the alias name, or the variable name. PROC REPORT Unwrapped: Exploring the Secrets behind One of the Most Popular Procedures in Base SAS ® Software, continued The following example code shows this concept: proc report nowd data=sashelp. class list; col age sex, (weight height total); define age / group; define sex / across; define heightweight/ sum format=8. 2; define total / computed format=8. 2; compute total; _c4_=_c 2_/_c3_; _c7_=_c5_/_c6_; endcomp; run; The COMPUTE TOTAL block shows two assignment statements. Each assignment corresponds to a column of WEIGHT, HEIGHT, and TOTAL for each unique value of the across variable SEX. The resulting output is shown in Output 3. Output 3. Output Using _Cn_ in the COMPUTE TOTAL Calculations A numeric input data set variable can also be defined as DISPLAY, GROUP, ORDER, or COMPUTED. Because there is no statistic associated with these definitions, the numeric input data set variable name is used in a COMPUTE block. Regardless of the definition, the numeric report-item can still be used in any computation. However, for GROUP or ORDER definitions, behind the scenes the values are evaluated from the printed output report instead of the input data. This means that if the ORDER or GROUP defined variable for a particular row and column shows as a blank on the printed output report, a blank is the value that will be used for any computation or evaluation. The following code shows three different methods for assigning the value of the ORDER variable AGE to a COMPUTED variable. proc report nowd data=sashelp. class; col age newage1 newage2 newage3; define age / order; define newage1 / computed; define newage2 / computed; define newage3 / computed; /* method 1 */ compute newage1; newage1=age*1. 5; endcomp; /* method 2 */ ompute newage2; if age ne . then hold_age=age; newage2=hold_age*1. 5; endcomp; /* method 3 */ compute before age; before_age=age; endcomp; compute newage3; newage3=before_age*1. 5; endcomp; run; 4 PROC REPORT Unwrapped: Exploring the Secrets behind One of the Most Popular Procedures in Base SAS ® Software, continued In the first method, the value for NEWAGE1 will contain a value only when AGE has a value for the sam e row. In the second method, the value of NEWAGE2 will contain a value for every row because it is obtaining a value from the DATA step variable HOLD_AGE. In the third method, the value of NEWAGE3 will contain a value for every row because it is obtaining a value from the DATA step variable BEFORE_AGE. The DATA step variable is created in the COMPUTE BEFORE AGE block. Behind the scenes, a DATA step variable changes values only through the code instructions. Also, behind the scenes, GROUP and ORDER numeric input data set variables are internally set to a blank in the printed output report at the RBREAK level. A COMPUTE AFTER block with an assignment statement for a numeric GROUP or ORDER variable at the RBREAK level will be ignored. A DISPLAY is always set to a blank at the BREAK and RBREAK levels. If you are routing the report output to an ODS destination, using a COMPUTE block CALL DEFINE statement with the STYLE attribute name and a style option that will accept text, such as PRETEXT=, is a way to override the blank values. A COLUMN STATEMENT VARIABLE VERSUS A DATA STEP VARIABLE PROC REPORT creates a column type of output report based on the variables and statistics listed in the COLUMN statement. Any variable from the input data set that is to be used as a report column or used in a COMPUTE block has to be listed in the COLUMN statement. The placement of the report items, variables, and statistics in the COLUMN statement is very important. PROC REPORT reads and processes the report items from the COLUMN statement in a left-to-right, top-to-bottom direction. Until the report item is processed, it will be initialized to missing for numeric variables and blank for character variables. Once the entire COLUMN statement report-items are processed for a row, PROC REPORT reinitializes all of the report-items back to missing for numeric and blank for character variables. Then PROC REPORT begins the process all over again for the next row of data by processing the report items in the COLUMN statement in a left-to-right direction. Behind the scenes, PROC REPORT consolidates all the input data set variables and statistics listed in the COLUMN statements for the execution of RBREAK BEFORE and BREAK BEFORE statements. For example, the RBREAK, meaning the report break, in the following code is calculated first: proc report nowd data=sashelp. class; col sex age,(height weight); define age / group; define height / min format=8. 2 Height min; efine weight / max format=8. 2 Weight max; rbreak before / summarize; run; The output is shown in Output 4. Output 4. PROC REPORT Output Showing the RBREAK Values COMPUTE blocks are also sensitive to the placement of the variables and statistics in the COLUMN statement. As PROC REPORT processes the report-items in a left-to-right direction, any associated COMPUTE blocks are also processed in the same order. This means th at in a COMPUTE block that is based on a COLUMN statement reportitem, any referenced variable or statistic to the right of the COMPUTE block variable is missing. Simply put, PROC REPORT does not know about any report-item that is to the right of the COMPUTE block variable in the COLUMN statement. A DATA step variable, also referred to as a temporary variable, is different from the COLUMN statement variable. A DATA step variable is created and used in a COMPUTE block. It is not part of the COLUMN statement. The value of the DATA step variable comes directly from the code in a COMPUTE block. DATA step variables are often used in IF statements when there is a comparison of the current row value to that of the value in the DATA step variable. PROC REPORT recomputes a COMPUTED variable value at every row, including at the BREAK and RBREAK rows. Values are not accumulated. An accumulated value can be calculated quickly using a DATA step variable in a 5 PROC REPORT Unwrapped: Exploring the Secrets behind One of the Most Popular Procedures in Base SAS ® Software, continued COMPUTE block because the value changes through the code only. Behind the scenes, DATA step variables used to accumulate values also include values at the BREAK and RBREAK levels. Adding an IF statement to check the value of the _BREAK_ automatic variable will help control when the accumulations takes place. In the following code, the computed variable TOTAL_AGE is the sum of two variables from the COLUMN statement. ACCUM_AGE is the accumulated value of AGE stored in the DATA step variable TEMP_AGE. proc report nowd data=sashelp. class; col age total_age accum_age height weight; define age / group; define height / min format=8. 2 Height min; define weight / max format=8. 2 Weight max; define total_age / computed; define accum_age / computed; compute total_age; if _break_ eq then total_age+age; endcomp; compute accum_age; if _break_ eq then temp_age+age; accum_age=temp_age; endcomp; break after / summarize; run; The output is shown in Output 5. Output 5. Comparison of the TOTAL_AGE Column and the ACCUM_AGE Column Notice the difference between the TOTAL_AGE column and the ACCUM_AGE column in Output 5. The TOTAL_AGE and AGE values are reinitialized for every row so that the values are not accumulated. The ACCUM_AGE and AGE values are reinitialized for every row but the TEMP_AGE value is not. TEMP_AGE is a DATA step variable and is not listed in the COLUMN statement. The result is an accumulated column for ACCUM_AGE. The _BREAK_ automatic variable will be blank for detail rows. A quick way to determine the value of a _BREAK_ variable value is to create an output data set with the OUT= option in the PROC REPORT statement and examine the _BREAK_ values in the output data set. DISCOVERING WAYS TO MOVE COLUMN HEADERS By default, the column heading values come from the label in the DEFINE statement. If you do not specifically specify a label in your code either in the DEFINE statement or through a LABEL statement, add the LIST option to the PROC REPORT statement, submit your code, and look at the code that is created in the SAS log. Behind the scenes, PROC REPORT will generate the default values it needs to create the output report. One of the default values is the label specified in the DEFINE statement. All of the column headings from the label option in the DEFINE statement span over a single column with one exception, variables that are defined as across variables. A column heading for an across variable can span over multiple columns. In the COLUMN statement, a comma after the across variable indicates which variable or group of variables are associated with the across variable. An example of PROC REPORT code containing an across variable is shown below: 6 PROC REPORT Unwrapped: Exploring the Secrets behind One of the Most Popular Procedures in Base SAS ® Software, continued title Default Column Headers; proc report nowd data=sashelp. shoes; column Region Product,Sales; define Region / group format= $25. Region; define Product / across format= $14. Product; define Sales / sum format= DOLLAR12. Total Sales; run; Output 6 shows the PROC REPORT example output. Output 6. Default Column Heading with an Across Label Spanning over Multiple Columns Behind the scenes, each unique value of an across variable is transposed from a column to a row. The row data is not available for any further processing within the code as it now becomes a column heading. In Output 6, each value of Product becomes a column with the Product value as the column heading. Under each Product column heading value is the Sales variable column heading and data for the particular Product value. The heading label Total Sales for every column is redundant. The output report would look better if Total Sales were removed from under the Product column heading and placed above the Product column headings. PROC REPORT provides a way to add column heading information that can span over multiple columns by using a SPANNED HEADER. The SPANNED HEADER is used in the COLUMN statement in this way: column (spanned header text variable-list)†¦; The following example code shows three different methods for using the DEFINE statement and SPANNED HEADERS for creating the column heading: proc report nowd data=sashelp. shoes split=*; olumn region ((1)Total Sales (1)Product (2)Total Sales*(2)Product product, sales); define region / group format= $25. Region; define product / across format= $14. (3)Total Sales (3)Product ; define sales / sum format=DOLLAR12. ; run; You can mix and match the methods. There is no best practice for using each method. The method that you choose depends on the look that you want for the column heading. The output is shown in Output 7. 7 PROC REPORT Unwrapped: Exploring the Secrets behind One of the Most Popular Procedures in Base SAS ® Software, continued Output 7. Moved Column Headings from Different Methods The three different methods are numbered in the example code and the output shown in Output 7: method (1) uses multiple SPANNED HEADER text; method (2) uses SPANNED HEADER text with the PROC REPORT SPLIT= character of * to force the text to continue on the next row; method (3) uses multiple labels in the DEFINE statement (you can also use a split character here). Let’s choose method (1) for the column heading and move the column heading to the top row. You can remove the label from the DEFINE statement by replacing the Region text with a blank â€Å" â€Å" and moving the Region text to a SPANNED HEADER in the COLUMN statement. There are three rows of headers. This means that the text of Region will need to be pushed up to the top row. You can do this by adding blank SPANNED HEADER text after the Region text in the COLUMN statement. Here is the modified PROC REPORT code with method (1) and the column heading text of Region: proc report nowd data=sashelp. shoes split=*; column (Region Region) (Total Sales Product Product , Sales); define Region / group format= $25. ; define Product / across format= $14. ; define Sales / sum format=DOLLAR12. ; run; Output 8 shows the output. Output 8. Moving Column Headings Using Blank SPANNED HEADERS Behind the scenes, when there is a blank header row and the output is routed to an ODS destination, the blank row is removed automatically. This does not affect the LISTING output. If you want to preserve the blank row, change the blank label on one of the DEFINE statements that is not an across variable to some value. Then add a style 8 PROC REPORT Unwrapped: Exploring the Secrets behind One of the Most Popular Procedures in Base SAS ® Software, continued statement for the header, assigning the foreground color to the background color. For example, if your column heading background is purple, then the style statement for the DEFINE statement would look something like this: style(header)=[background=purple foreground=purple] With the background and the foreground assigned to the same color, any text in the label will blend into the background color. CHANGING DEFAULT ATTRIBUTES WITH STYLES Beginning with SAS 9. 3, the default output destination is HTML. Behind the scenes, PROC REPORT is using the HTMLBLUE style. All the output in this paper all uses this default destination. What if you are not fond of the HTMLBLUE style? Then, what do you do if you want to change the default style of your output report? If you want to change the style of HTMLBLUE to another style that is supplied in the Sashelp. Tmplmst template store, you can run the following code to create a list of all the styles that are available: proc template; list styles; run; You can apply the styles by adding an ODS statement with the specified style before the PROC REPORT statement. For example, if you want to use the FESTIVAL style instead of the default HTMLBLUE style, the ODS statement would look similar to this: ods html style=festival; PROC REPORT also provides the ability to change the styles of the different report locations. Here are the style location values and a description for each that indicates which part of the report is affected: ? ? ? ? ? ? REPORT—the report as a whole HEADER|HDR—the column headings COLUMN—the column cells LINES—the lines generated by LINE statements SUMMARY—the summary rows created from BREAK and RBREAK statements CALLDEF—the cells identified by a CALL DEFINE statement All of the style locations are valid in the PROC REPORT statement. These styles apply to the entire location that is specified. The style locations can also be combined if the same attribute is being applied to multiple locations. This is the correct syntax: style= The following code shows how to apply the styles in the PROC REPORT statement: ods html style=festival; title Styles on the PROC REPORT statement; proc report nowd data=sashelp. class(obs=5) split=* style(report)=[outputwidth=7in] style(column)=[background=lavender] style(header)=[foreground=green] style(summary)=[background=purple foreground=white] style(lines)=[background=lime] style(calldef)=[background=yellow foreground=black]; olumn name age sex weight height; define name / display; define age / order; define sex / display; define heightweight / sum; break after age / summarize; rbreak after / summarize; compute before; line this is the beginning; endcomp; 9 PROC REPORT Unwrapped: Exploring the Secrets behind One of the Most Popular Procedures in Base SAS ® Software, continued compute age; if _break_ ne then call define(age,style,style=[pr etext=total]); endcomp; run; The STYLE options in the preceding PROC REPORT statement are formatting the output in this way: ? ? ? style(report) sets the report output width to 7 inches. style(column) sets the background for all of the columns to lavender. style(header) applies a green foreground to all of the headers. style(summary) sets all of the summary rows created from BREAK and RBREAK statements with a ? ? style(lines) sets the line statements to a background of lime. style(calldef) sets the foreground to black and background to yellow for the CALL DEFINE locations. background of purple and a foreground of white. The resulting report output is shown in Output 9. Output 9. Changing Default Styles in the PROC REPORT Statement The DEFINE statement supports two types of styles: STYLE(COLUMN) and STYLE(HEADER). STYLE(COLUMN) applies to the entire column but will not override any styles that are applied to other locations in the column. Using the same code in this section, you can modify the DEFINE statement for the NAME variable that creates the Name column like this: define name / display style(column header)=[background=plum]; The background of the HEADER and COLUMN locations for the NAME variable is set to plum. Because styles were applied already to the SUMMARY location, only the header and detail cells for the NAME column are changed to plum. A CALL DEFINE statement is used to override the SUMMARY style for the NAME column. The CALL DEFINE statement is discussed more in the next section. Output 10 is the resulting report output. Output 10. Changing the Default Styles for the NAME Column Using a DEFINE Statement 10 PROC REPORT Unwrapped: Exploring the Secrets behind One of the Most Popular Procedures in Base SAS ® Software, continued The BREAK and RBREAK statements support style changes for summary lines, customized lines, or both. A summary line is created from the BREAK or RBREAK statements. A customized line is created from a LINE statement within a COMPUTE BEFORE or a COMPUTE AFTER COMPUTE block. The is a break-variable that is defined as either GROUP or ORDER or the _PAGE_ location. A style on the BREAK and RBREAK statements will not override a cell style that is created by a CALL DEFINE statement or the STYLE(CALLDEF) option in the PROC REPORT statement. A CALL DEFINE statement will be used to make the style changes in this case. Using the same code in this section, you can modify the RBREAK statement like this: break after / summarize style=[background=pink foreground=black font_weight=bold]; The COMPUTE BEFORE or a COMPUTE AFTER supports a style option in the COMPUTE statement. A forward slash ‘/’ precedes the style option in the COMPUTE statement. The style option only applies to the LINE statement and will override any previous STYLE(LINES) requests. The style applies to all of the LINE s tatements within the COMPUTE block. Using the code from this section, a COMPUTE AFTER AGE block is added to show a style modification to the foreground of the LINE statement output. ompute after age/ style=[foreground=red]; line this is after age; endcomp; A CALL DEFINE is a statement within a COMPUTE block. To change a style using a CALL DEFINE statement, the STYLE attribute is specified for the attribute-name and the style option is specified as the value. The following is the syntax for a CALL DEFINE statement: call define (column-id | _ROW_ , attribute-name, value); Here is the code with all of the style modifications: ods html style=festival; title Changing Default Attributes with Styles; proc report nowd data=sashelp. lass(obs=5) split=* style(report)=[outputwidth=7in] style(column)=[background=lavender] style(header)=[foreground=green] style(summary)=[background=purple foreground=white] style(lines)=[background=lime] style(calldef)=[background=yellow foreground=black]; colum n name age sex weight height; define name / display style(column header)=[background=plum]; define age / order; define sex / display; define heightweight / sum; break after age / summarize; rbreak after / summarize style=[background=pink foreground=black font_weight=bold]; ompute before; line this is the beginning; endcomp; compute age; if _break_ ne then call define(age,style,style=[pretext=total]); endcomp; compute after age/ style=[foreground=red]; line this is after age; endcomp; run; The updated output is shown in Output 11. 11 PROC REPORT Unwrapped: Exploring the Secrets behind One of the Most Popular Procedures in Base SAS ® Software, continued Output 11. Final Report Output with Changes to Default Attributes Using Style Options You also can change styles by using inline formatting. Inline formatting is a feature of the Output Delivery System that enables you to insert simple formatting text into ODS output by using the ODS ESCAPECHAR statement. For example, here is a TITLE statement and the resulting output: title This is ^{style [color=red font_weight=bold] RED}; This is RED The inline formatting in the TITLE statement changes the text of RED to the color of red. The caret (^) in the TITLE statement is the declared ODS ESCAPECHAR. The ODS ESCAPECHAR statement has to be submitted before any inline formatting will take place. The caret (^) can be any unique character that would not normally be in your code. USING THE CALL DEFINE STATEMENT The previous section discussed using the CALL DEFINE statement as a way to change a style by specifying the STYLE attribute for the attribute-name and the STYLE= op