SAS 9.1 Output Delivery System Users Guide

PROC TEMPLATE ;

• EDIT definition- path -1 <AS definition-path-2 > < / STORE= libref . template-store >;

  • statements-and-attributes

  • END ;

• DEFINE COLUMN column-path < / STORE= libref . template-store >;

  • statements-and-attributes

  • END ;

• DEFINE FOOTER footer-path < / STORE= libref . template-store >

  • statements-and-attributes

  • END ;

• DEFINE HEADER definition- name ;

  • statements-and-attributes

  • END ;

• DEFINE TABLE table-path </ STORE= libref . template-store >;

  • statements-and-attributes

  • END ;

The following table lists the statements that you use to add different features to your SAS tabular output.

EDIT Statement

Edits an existing definition for a table, column, header, or footer

Requirement: An END statement must follow the EDIT statement, after all of the editing instructions.

Interaction: In some cases, you can use an EDIT statement inside a set of editing instructions.

• When you edit a table definition, you can also edit one or more column, header, or footer definitions that are defined in the table.

• When you edit a column definition, you can also edit one or more header definitions that are defined for that column.

Restriction: If you edit a definition that is a link, you break the link and create a separate definition.

Featured in: Example 1 on page 515

• EDIT definition-path-1 < AS definition-path-2 > </ STORE= libref . template-store >;

  • attribute-statements ;

  • END ;

Required Arguments

definition-path-1

• specifies a definition to edit. definition-path-1 consists of one or more names , separated by periods. Each name represents a directory in a template store. (A template store is a type of SAS file.)

• Interaction: The STORE= option specifies a particular template store to read from and to write to.

• Tip: You can determine what definitions a procedure or DATA step uses by submitting the ODS TRACE ON statement before you run the SAS program (see ODS TRACE Statement on page 197).

Options

AS definition-path-2

• specifies the location in which to store the edited definition, where definition-path-2 consists of one or more names, separated by periods. Each name represents a directory in a template store. (A template store is a type of SAS file.) By default, PROC TEMPLATE writes the edited definition to the first template store that you can write to in the current path.

• Default: If you do not specify AS definition-path-2 , PROC TEMPLATE writes the edited definition to definition-path-1 in the first template store that you can write to.

• Restriction: If the current EDIT statement is inside a set of editing instructions, then you cannot use the AS definition-path-2 option.

STORE= libref . template-store

• specifies the template store from which to read definition-path-1 and in which to store definition-path-2 .

Statements and Attributes

The EDIT statement supports the same statements and attributes as the DEFINE TABLE statements. For more information, see DEFINE TABLE Statement on page 410.

Editing an Existing Definition

There are two steps to follow when you edit an existing definition.

1. Open a copy of the specified file.

   By default, PROC TEMPLATE looks for definition-path-1 in the list of template stores that is defined by the PATH statement (see PATH Statement on page 276). It opens a copy of the first definition path that it finds in a template store that has read access.

2. Save the modified file.

   PROC TEMPLATE writes the modified definition to the first template store in the current path with update access. If you do not specify a second definition path to write to, then it uses definition-path-1 . Therefore, if you have update access to the template store from which you read definition-path-1 , then you actually modify the original definition. Otherwise , the modified file is written to a template store to which you do have update access.

   If you do specify a second definition path, then PROC TEMPLATE writes the edited definition to the specified path in the first template store to which you have write access.

DEFINE COLUMN Statement

Creates a definition for a column

Requirement: An END statement must be the last statement in the definition.

Interaction: A column definition can include one or more header definitions.

See also: DEFINE HEADER Statement on page 395

Featured in: Example 3 on page 528 and Example 5 on page 539

• DEFINE COLUMN column-path < / STORE= libref . template-store >;

  • < column-attribute-1 ;< column-attribute-n ;>>

  • CELLSTYLE expression-1 AS < style-element-name ><[ style-attribute-specification(s) ] > < , expression-n AS < style-element-name ><[ style-attribute-specification(s) ]>>;

  • COMPUTE AS expression ;

  • DEFINE HEADER definition-path ;

    • statements-and-attributes

    • END ;

  • DYNAMIC variable-1 < text-1 > < variable-n < text-n >>;

  • MVAR variable-1 < text-1 > < variable-n < text-n >>;

  • NMVAR variable-1 < text-1 > < variable-n < text-n >>;

  • NOTES text ;

  • TRANSLATE expression-1 INTO expression-2 < , expression-n INTO expression-m >;

  • END ;

Required Arguments

column-path

• specifies where to store the column definition. A column-path consists of one or more names, separated by periods. Each name represents a directory in a template store. (A template store is a type of SAS file.) PROC TEMPLATE writes the definition to the first template store that you can write to in the current path.

  Restriction: If the definition is nested inside another definition, definition-path must be a single-level name because the nested definition is stored where the original definition is stored.

  Restriction: If you want to reference the definition that you are creating from another definition, do not nest the definition inside another one. For example, if you want to reference a column definition from multiple tables, do not define the column inside a table definition.

Options

STORE= libref . template-store

• specifies the template store in which to store the definition. If the template store does not exist, it is created.

• Restriction: If the definition is nested inside another definition, you cannot use the STORE= option for the nested definition because it is stored where the original definition is stored.

• Restriction: The STORE= option does not become part of the definition.

Column Attributes

This section lists all the attributes that you can use in a column definition. For all attributes that support a value of ON, the following forms are equivalent:

ATTRIBUTE-NAME ATTRIBUTE-NAME=ON

For all of the attributes that support a value of variable , variable can be any variable that you declare in the column definition with the DYNAMIC, MVAR, or NMVAR statement. If the attribute is a boolean, then the value of variable should resolve to either true or false as shown in the following table:

BLANK_DUPS<=ON | OFF | variable >

• specifies whether or not to suppress the value of a variable from one row to the next, if the value does not change based on the formatted value of the variable.

• Default: OFF

• Interaction: If the CLASSLEVELS= table attribute is in effect, ODS ignores BLANK_DUPS=ON when any value changes in a preceding column that is also marked with BLANK_DUPS=ON.

• ODS Destinations: All except OUTPUT. Note that when the PRINTER destination suppresses the value of a variable, it also suppresses the horizontal rule above the blank cell .

• Featured in: Example 5 on page 539

BLANK_INTERNAL_DUPS<=ON | OFF | variable >

• specifies whether or not to suppress the value of a variable from one row to the next, if the value does not change based on the raw value of the variable.

• Default: OFF

• Interaction: If the CLASSLEVELS= table attribute is in effect, ODS ignores BLANK_INTERNAL_DUPS=ON when any value changes in a preceding column that is also marked with BLANK_INTERNAL_DUPS=ON.

• ODS Destinations: All except OUTPUT. Note that when the PRINTER destination suppresses the value of a variable, it also suppresses the horizontal rule above the blank cell.

CHOOSE_FORMAT= COMPROMISE | MAX | MAX_ABS | MIN_MAX

• selects a format based on the actual values in the column of the table.

• Default: If you omit the CHOOSE_FORMAT column attribute, then the default format is either determined by the data component or by other attributes.

• Restriction: CHOOSE_FORMAT is not supported for computed columns because those columns values are computed outside of the data object .

• Tip: If you specify a small value for the FORMAT_WIDTH= option, then CHOOSE_FORMAT may create a dw.3 format.

• ODS Destinations: All

• See: For more information about column formats, see How Are Values in Table Columns Formatted? on page 514.

• COMPROMISE

  • looks at all of the values in the column and selects a good compromise format to work well for most values, but extreme values may shift to BEST format.

  • Tip: FORMAT_NDEC=d specifies the precision in digits.

  • Tip: The FORMAT_WIDTH= option suggests a maximum width. The actual format width may be smaller or it may be larger.

• MAX

  • selects a format based on the maximum value in the column. Values are all expected to be positive so no space is reserved for a minus sign.

  • Default: By default, FORMAT_WIDTH=10 and FORMAT_NDEC= is ignored.

• MAX_ABS

  • selects a format based on the maximum absolute value in the column. The format reserves space for a minus sign whether it is needed or not.

• MIN_MAX

  • selects a format based on the minimum and maximum value in the column. The format reserves space for a minus sign only where it is actually needed.

  • Interaction: If FORMAT_NDEC= d is specified, a maximum of d decimal places is used.

DATA_FORMAT_OVERRIDE<=ON | OFF | variable >

• specifies which format to use if both a column definition and a data component specify a format.

• Default: OFF

• ODS Destinations: All

• ON

  • Uses the format in the data component.

• OFF

  • Uses the format in the column definition.

• variable

  • Uses the format of the specified variable.

DATANAME= column-name

• specifies the name of the column in the data component to associate with the current column.

• Default: By default, ODS associates the current column with a column of the same name in the data component.

• ODS Destinations: All

DEF_SPLIT

• specifies which special characters in headers for generic columns are to be used as split characters.

• ODS Destinations: All

DROP<=ON | OFF | variable >

• specifies whether or not to include the column in an output data set.

• Default: OFF

• ODS Destinations: OUTPUT

FLOW<=ON | OFF | variable >

• specifies whether or not to wrap the text in the current column if it is too long to fit in the space that is provided.

• Default: ON if the format width of the column is greater than the column width. OFF if the format width of the column is not greater than the column width.

• See also: MAXIMIZE= on page 384

• ODS Destinations: LISTING

  Note: The HTML and PRINTER destinations always wrap the text if it is too long to fit in the space that is provided.

FORMAT= format-name < format-width < decimal-width >> | variable

• specifies the format for the column.

• Default: If you do not specify the FORMAT= , PROC TEMPLATE uses the format that the data component provides. If the data component does not provide a format, PROC TEMPLATE uses

  • BEST8. for integers

  • 12.3 for doubles

  • the length of the variable for character variables.

• Restriction: If you specify a format width for a numeric column, then its value cannot exceed 32.

• ODS Destinations: All

FORMAT_NDEC= number | variable

• specifies the number of decimals for the column.

• Default: the decimal width that is specified with the FORMAT= column attribute.

• Range: Number is a whole number from 0 to 32

• Interaction: If you specify a decimal width using both the FORMAT= and the FORMAT_NDEC= attributes, then PROC TEMPLATE uses the width that you specify with the FORMAT= attribute.

• ODS Destinations: All

FORMAT_WIDTH= positive-integer | variable

• specifies the format width for the column.

• Default: If you omit the column attribute FORMAT_WIDTH=, then the format specified in the FORMAT= attribute is used.

• Range: 1 to 32 for numeric variables; operating system limit for character variables

• Interaction: If you specify a format width using both the FORMAT= and the FORMAT_WIDTH= attributes, then PROC TEMPLATE uses the width that you specify with the FORMAT= attribute.

• ODS Destinations: All

FUZZ= number | variable

• supplies a numeric value against which values in the column are compared to eliminate trivial values from printing. A number whose absolute value is less than or equal to the FUZZ= value is printed as 0. However, the real value of the number is used in any computations based on that number.

• Default: the smallest representable floating-point number on the computer that you are using

• ODS Destinations: All except OUTPUT

GENERIC<=ON | OFF | variable >

• specifies whether or not the column definition can be used by more than one column. Generic columns are useful in tables with many similar columns. For example, the table definitions for both PROC SQL and the DATA step define only two columns: one for character variables and one for numeric variables. When a program runs, it determines which column definition the data component should use for each column.

• Default: OFF

• ODS Destinations: All except OUTPUT

• Featured in: Example 3 on page 528 and Example 5 on page 539

GLUE= integer | variable

• Influences the places at which ODS splits a table when it creates multiple data panels. ODS creates multiple data panels from a table that is too wide to fit in the allotted space. The higher the value of GLUE= is, the less likely it is that ODS will split the table between the current column and the column to its right.

• Default: 1

• Range: -1 to 327

• Tip: A value of -1 forces the table to split between the current column and the column to its right.

• ODS Destinations: LISTING, PRINTER family, and RTF

HEADER= header-specification

• specifies the text for the column header. header-specification can be one of the following:

• text

  • specifies the actual text of the header.

  • Requirement: text must be enclosed by quotation marks.

• header-name

  • specifies the name of a header definition to use. You create a header definition with the DEFINE HEADER statement (see DEFINE HEADER Statement on page 395). If header-name is a single-level name, the header definition must occur within the current column definition.

• variable

  • specifies the name of a variable that you declare with the DYNAMIC, MVAR, or NMVAR statement. The value of the variable becomes the column header.

• _LABEL_

  • Uses the label that is specified in the data component for the column header.

• Default: _LABEL_

• Interaction: If you are using the OUTPUT destination, then the HEADER= attribute does not change the label of the variable in the data set. To change the label in the data set, you must use the LABEL= attribute.

• Tip: The HEADER= option provides a simple way for you to specify the text of a column header. If you want to customize the header further, use the DEFINE HEADER statement with the appropriate header attributes. (See DEFINE HEADER Statement on page 395.)

• Tip: You can use the split character in the text of the header to force the text to a new line.

• See also: LABEL= on page 384 and TEXT_SPLIT= on page 387.

• ODS Destinations: All

• Featured in: Example 3 on page 528 and Example 1 on page 342

ID<=ON | OFF | variable >

• specifies whether or not the column is an ID column. An ID column is repeated on each data panel. (ODS creates multiple data panels when a table is too wide to fit in the allotted space.)

• Default: OFF

• Tip: ODS treats all columns up to and including a column that is marked with ID=ON as ID columns.

• ODS Destinations: LISTING and PRINTER family

• Featured in: Example 3 on page 528

JUST= justification | variable

• specifies the horizontal justification of the format field within the column (and of the header if the definition for the header does not include JUST=).

  • justification can be one of the following:

• LEFT

  • specifies left justification.

  • Alias: L

• RIGHT

  • specifies right justification.

  • Alias: R

• CENTER

  • specifies center justification.

  • Alias: C

• Default: LEFT for columns that contain character values; RIGHT for columns that contain numeric values.

• Interaction: For the Listing destination, ODS justifies the format field within the column width. At times, you may need to specify the JUSTIFY= attribute to get the results that you want. See the discussion of JUSTIFY= on page 384.

• See also: FORMAT= on page 381 and WIDTH= on page 388.

• Main discussion: How Are Values in Table Columns Justified? on page 512

• ODS Destinations: All except OUTPUT

• Featured in: Example 1 on page 515

JUSTIFY<=ON | OFF | variable >

• specifies whether to justify the format field within the column or to justify the value within the column without regard to the format field.

• Default: OFF

• Interaction: JUSTIFY=ON can interfere with decimal alignment.

• Tip: If you translate numeric data to character data, you may need to use JUSTIFY= to align the data.

• Main discussion: How Are Values in Table Columns Justified? on page 512

• Featured in: Example 5 on page 539

  ODS Destinations: All destinations except LISTING behave as if JUSTIFY=ON

LABEL= text | variable

• specifies a label for the column in the output data set.

• Default: If you do not specify a label, ODS uses the label that is specified in the data component. If no label is specified in the data component, ODS uses the header for the column as the label.

• ODS Destinations: OUTPUT

• Tip: If the Output destination is open, LABEL= provides a label for the corresponding variable in the output data set. This label overrides any label that is specified in the data component.

MAXIMIZE<=O N | OFF | variable >

• specifies whether to try to divide the text equally among all lines or to maximize the amount of text in each line when the text in the column uses more than one line. For example, if the text spans three lines, MAXIMIZE=ON might result in 45% of the text on the first line, 45% of the text on the second line, and 10% of the text on the third line. MAXIMIZE=OFF would result in 33% of the text on each line. MAXIMIZE=ON may write lines of text that vary greatly in length. MAXIMIZE=OFF may result in using less than the full column width.

• Default: OFF

• Interaction: This attribute is effective only if the column is defined with FLOW=ON (see the discussion of FLOW= on page 381).

• ODS Destinations: LISTING

MERGE<=ON | OFF | variable >

• specifies whether or not to merge the current column with the column immediately to its right. When you set MERGE=ON for the current column, the data in each row of the column is merged with the data in the same row of the next column. ODS applies the format, justification, spacing, and prespacing attributes to each column independently. Then, it concatenates the columns. Finally, it applies to the concatenated data all the remaining attributes that are specified on the column that does not have MERGE= set.

• Default: OFF

• Restriction: You cannot use both MERGE=ON and PRE_MERGE=ON in the same column definition. You cannot merge or premerge a column with another column that has either MERGE=ON or PRE_MERGE=ON. Note that you can merge three columns by setting MERGE=ON for the first column, no merge or premerge attributes for the second column, and PRE_MERGE=ON for the third column.

• See also: PRE_MERGE= on page 385

• ODS Destinations: All except OUTPUT

OPTIONAL<=ON | OFF | variable >

• specifies whether or not to delete the current column from the output object if doing so enables all the remaining columns to fit in the space that is provided without splitting the table into multiple data panels.

• Default: OFF

• Interaction: If multiple column definitions contain OPTIONAL=ON, PROC TEMPLATE includes either all or none of these columns in the output object.

• ODS Destinations: LISTING

OVERLINE<=ON | OFF | variable >

• specifies whether or not to draw a continuous line in the current column above the first table footer (or, if there is no table footer, below the last row of the column). PROC TEMPLATE uses the second formatting character to draw the line. (See the discussion of FORMCHAR= on page 417.)

• Default: OFF

• ODS Destinations: LISTING

PARENT= column-path

• specifies the column definition that the current definition inherits from. A column-path consists of one or more names, separated by periods. Each name represents a directory in a template store. (A template store is a type of SAS file.) The current definition inherits from the specified column in the first template store that you can read from in the current path.

  When you specify a parent, all the attributes and statements that are specified in the parent s definition are used in the current definition unless the current definition specifically overrides them.

• ODS Destinations: All

PREFORMATTED<=ON | OFF | variable >

• specifies whether or not to treat the text as preformatted text. When text is preformatted, ODS honors line breaks as well as leading, trailing, and internal spaces. It also displays the text in a monospace font.

• Default: OFF

• Interaction: When PREFORMATTED=ON, ODS uses the datafixed style element unless you specify another style element with the STYLE= column attribute.

• ODS Destinations: HTML, MARKUP family, PRINTER family, and RTF

PRE_MERGE<=ON | OFF | variable >

• specifies whether or not to merge the current column with the column immediately to its left. When you set PRE_MERGE=ON for the current column, the data in each row of the column is merged with the data in the same row of the previous column. ODS applies the format, justification, spacing, and prespacing attributes to each column independently. Then, it concatenates the columns. Finally, it applies to the concatenated data all the remaining attributes that are specified on the column that does not have PRE_MERGE= set.

• Default: OFF

• Restriction: You cannot use both MERGE=ON and PRE_MERGE=ON in the same column definition. You cannot merge or premerge a column with another column that has either MERGE=ON or PRE_MERGE=ON. Note that you can merge three columns by setting MERGE=ON for the first column, no merge or premerge attributes for the second column, and PRE_MERGE=ON for the third column.

• See also: MERGE= on page 384

• ODS Destinations: All except OUTPUT

PRE_SPACE= non-negative-integer

• specifies the number of blank characters to leave between the current column and the column immediately to its left.

• Default: A value in the range that is bounded by the COL_SPACE_MIN and COL_SPACE_MAX table attributes.

• Interaction: If PRE_SPACE= and SPACE= are specified for the same intercolumn space, ODS honors PRE_SPACE=.

• See also: SPACE= on page 386, COL_SPACE_MIN= on page 416, and COL_SPACE_MAX= on page 416

• ODS Destinations: LISTING

PRINT<=ON | OFF | variable >

• specifies whether or not to print the column.

• Default: ON

• See also: OPTIONAL= on page 385 and DROP= on page 381

• ODS Destinations: All except OUTPUT

PRINT_HEADERS<=ON | OFF | variable >

• specifies whether or not to print the column header and any underlining and overlining.

• Default: ON

• See also: UNDERLINE= on page 387 and OVERLINE= on page 385

• ODS Destinations: All except OUTPUT

SEPARATOR= character | variable

• specifies a separator character to append to each value in the column.

• Default: None

• Restriction: The SEPARATOR= column attribute is valid only for character variables.

• Tip: To specify a hexadecimal character as the separator character, put an x after the closing quote. For example, the following option assigns the hexadecimal character 2D as the separator character: separator= 2D x

• ODS Destinations: LISTING

SPACE= positive-integer | variable

• specifies the number of blank characters to leave between the current column and the column immediately to its right.

• Default: A value in the range that is bounded by the COL_SPACE_MIN and COL_SPACE_MAX table attributes.

• Interaction: If PRE_SPACE= and SPACE= are specified for the same intercolumn space, ODS honors PRE_SPACE=.

• See also: PRE_SPACE= on page 386, COL_SPACE_MIN= on page 416, and COL_SPACE_MAX= on page 416

• ODS Destinations: LISTING

STYLE=< style-element-name ><[ style-attribute-specification(s) ]>

• specifies the style element and any changes to its attributes to use for the current column. Neither style-attribute-specification nor style-element-name is required. However, you must use at least one of them.

  Note: You can use braces ({ and }) instead of square brackets ([ and ]).

  style-element-name

  • is the name of the style element to use to display the data in the column. The style element must be part of a style definition that is registered with the Output Delivery System. SAS provides some style definitions. You can create your own style definitions with PROC TEMPLATE (see DEFINE STYLE Statement on page 288). By default, ODS displays different parts of ODS output with different style elements. For example, by default, the data in a column is displayed with the style element data . The style elements that you would be most likely to use with the STYLE= column attribute are

    • data

    • datafixed

    • dataempty

    • dataemphasis

    • dataemphasisfixed

    • datastrong

    • datastrongfixed.

  • The style element provides the basis for displaying the column. Additional style attributes that you provide can modify the display.

  • For information on viewing a style definition so that you can see the style elements that are available, see Viewing the Contents of a Style Definition on page 319. For information about the default style definition that ODS uses, see The Default Style Definition for HTML and Markup Languages on page 320.

    style-element-name can be either the name of a style element or a variable whose value is a style element.

  • Default: data

• style-attribute-specification

  • describes the style attribute to change. Each style-attribute-specification has this general form:

    • style-attribute-name = style-attribute-value

  • For information on the style attributes that you can specify, see Style Attributes and Their Values on page 292.

• ODS Destinations: HTML, MARKUP family, PRINTER family, and RTF

• Featured in: Example 3 on page 528

• Tip: If you use the STYLE= attribute inside a quoted string, then you must add a space before or after the carriage return to prevent errors. SAS does not interpret a carriage return as a space. You must explicitly specify spaces in your quoted strings.

TEXT_SPLIT= character | variable

• specifies the split character for the data in the column. PROC TEMPLATE breaks a value in the column when it reaches that character and continues the value on the next line. The split character itself is not part of the data and does not appear in the column.

• Default: None

• ODS Destinations: All except OUTPUT

UNDERLINE<=ON | OFF | variable >

• specifies whether or not to draw a continuous line in the current column below the column header (or, if there is no column header, above the first row of the column). PROC TEMPLATE uses the second formatting character to draw the line.

• Default: OFF

• Main discussion: See the discussion of FORMCHAR= on page 417.

• ODS Destinations: LISTING

VARNAME= variable-name | variable

• specifies the name to use for the corresponding variable in an output data set.

• Default: If you do not specify VARNAME=, PROC TEMPLATE uses the value of the DATANAME= attribute. If you do not specify DATANAME=, PROC TEMPLATE uses the name of the column.

• Tip: If you use VARNAME= to specify the same name for different columns, a number is appended to the name each time that the name is used.

• ODS Destinations: OUTPUT

VJUST= justification | variable

• Specifies the vertical justification for the column. justification can be one of the following:

• TOP

  • places the first line of text as high as possible.

  • Alias: T

• CENTER

  • centers the text vertically.

  • Alias: C

• BOTTOM

  • places the last line of text as low as possible.

  • Alias: B

• Default: TOP for the Printer destination; CENTER for the HTML destination

• ODS Destinations: HTML, MARKUP family, PRINTER family, and RTF

• Featured in: Example 3 on page 528

WIDTH= positive-integer | variable

• specifies the width of the column in characters.

• Default: If you do not specify a width, PROC TEMPLATE uses the format width. If the column has no format associated with it, PROC TEMPLATE uses a width of

  • 8 for integers

  • 12 for doubles

  • data length for character variables.

• Interaction: The length of the column header can influence the width of the column.

• See also: WIDTH_MAX on page 388 and WIDTH= on page 406.

• ODS Destinations: LISTING

WIDTH_MAX= positive-integer | variable

• specifies the maximum width allowed for this column. By default, PROC TEMPLATE extends the width of the column if the header is wider than the data. The width of the column can be anywhere between the values of WIDTH= and WIDTH_MAX=.

• Default: the width of the format for the column

• ODS Destinations: LISTING

CELLSTYLE-AS Statement

Sets the style element of the cells in the column according to the values of the variables. Use this statement to set the presentation characteristics (such as foreground color , font face, flyover) of individual cells

Featured in: Example 5 on page 539

CELLSTYLE expression-1 AS < style-element-name ><[ style-attribute-specification(s) ] >< , expression-n AS < style-element-name ><[ style-attribute-specification(s) ]>>;

Required Arguments

expression

• is an expression that is evaluated for each cell in the column. It can be any expression that is valid in the WHERE statement (or the WHERE= data set option). For information on expressions that you can use in the WHERE statement, see Statements in SAS Language Reference: Dictionary . Use _VAL_ to represent the value of the current column. You may also reference symbols that you declared in a DYNAMIC, MVAR, or NMVAR statement in the definition.

  If expression resolves to TRUE (a non-zero value), the style element that is specified is used for the current cell. If expression is FALSE (zero), the next expression in the statement is evaluated. Thus, you can string multiple expressions together to format cells conditionally.

• Restriction: You can not reference the values of other columns in expression .

• Tip: Using an expression of 1 as the last expression in the CELLSTYLE-AS statement sets the style element for any cells that did not meet an earlier condition.

Options

Note: Neither style-attribute-specification nor style-element-name is required. However, you must use at least one of them.

style-attribute-specification

• describes a style attribute to set. Each style-attribute-specification has this general form:

  • style-attribute-name = style-attribute-value

• For information on the style attributes that you can set in a column definition, see Style Attributes and Their Values on page 292.

• Default: If you don t specify any style attributes to modify, ODS uses the unmodified style-element-name .

style-element-name

• is the name of the style element that is used to display the data in the column. The style element must be part of a style definition that is registered with the Output Delivery System. SAS provides some style definitions. You can create your own style definitions by using PROC TEMPLATE (see DEFINE STYLE Statement on page 288). By default, ODS displays different parts of ODS output with different style elements. For example, by default, the data in a column is displayed with the style element data . The style elements that you would be most likely to use with the CELLSTYLE-AS statement in a column definition are the following.

  • data

  • datafixed

  • dataempty

  • dataemphasis

  • dataemphasisfixed

  • datastrong

  • datastrongfixed.

• The style element provides the basis for displaying the column. Additional style attributes that you provide can modify the display.

• Default: data

• See also: Viewing the Contents of a Style Definition on page 319.

• See also: The Default Style Definition for HTML and Markup Languages on page 320.

COMPUTE AS Statement

Computes values for a column that is not in the data component, or modifies the values of a column that is in the data component

COMPUTE AS expression ;

Required Arguments

expression

• is an expression that assigns a value to each table cell in the column. It can be any expression that is valid in the WHERE statement (or the WHERE= data set option). For information on expressions that you can use in the WHERE statement, see Statements in SAS Language Reference: Dictionary .

  To reference another column in a COMPUTE-AS statement, use the name of the column. You can also reference symbols that you declared in a DYNAMIC, MVAR, or NMVAR statement in the current definition. In addition, if the column has values in the data component, you can reference the column itself in the expression. However, if you are creating a column that does not exist in the data component, you cannot reference the column in the expression because there is no underlying value to use.

  For example, the following DEFINE COLUMN block defines a column that contains the square root of the value in the column called source :

  define column sqroot; compute as sqrt(source); header='Square Root'; format=6.4; end;

• Tip: The COMPUTE AS statement can alter values in an output object. None of the definitions that SAS provides modifies any values. If you want to determine if a definition was provided by SAS, use the ODS VERIFY Statement on page 205. If the definition is not from SAS, the ODS VERIFY statement returns a warning when it runs the SAS program that uses the definition. If you receive such a warning, you can use the SOURCE statement to look at the definition and determine if the COMPUTE AS statement is used to alter values. (See SOURCE Statement on page 277.)

• Tip: Because you can use column names in expression , _VAL_ is not recognized as an alias for the current column.

DEFINE HEADER Statement

Creates a definition for a header inside a column definition

Main discussion: DEFINE HEADER Statement on page 395

DEFINE HEADER definition-name ;

• statements-and-attributes

• END ;

Required Arguments

definition-name

• specifies the name of the new header.

• Restriction: definition-name must be a single-level name.

  • Note: If you want to reference the header definition that you are creating from another definition, you must create it outside the column definition.

statements-and-attributes

• specify the statements and header attributes that you can use to define a header inside a column.

• See: DEFINE HEADER Statement on page 395

DYNAMIC Statement

Defines a symbol that references a value that the data component supplies from the procedure or DATA step

Scope: You can use the DYNAMIC statement in the definition of a table, column, header, or footer. A dynamic variable that is defined in a definition is available to that definition and to all the definitions that it contains.

Featured in: Example 1 on page 342 and Example 2 on page 348

DYNAMIC variable-1 < text-1 > < variable-n < text-n >>;

Required Arguments

variable

• Names a variable that the data component supplies. ODS resolves the value of the variable when it binds the definition and the data component.

• Tip: Dynamic variables are most useful to the authors of SAS procedures and to DATA step programmers.

Options

text

• is text that you can place in the definition to explain the dynamic variable s use. Text of this type becomes part of the compiled definition, which you can view with the SOURCE statement, whereas SAS comments do not.

MVAR Statement

Defines a symbol that references a macro variable. ODS will use the value of the variable as a string. References to the macro variable are resolved when ODS binds the definition and the data component to produce an output object

Scope: You can use the MVAR statement in the definition of a table, column, header, or footer. A macro variable that is defined in a definition is available to that definition and to all the definitions that it contains.

Featured in: Example 3 on page 528 and Example 1 on page 342

MVAR variable-1 < text-1 > < variable-n < text-n >>;

Required Arguments

variable

• Names a macro variable to reference in the definition. ODS will use the value of the macro variable as a string. ODS does not resolve the value of the macro variable until it binds the definition and the data component.

• Tip: You must declare macro variables this way in a definition. For example, to use the automatic macro variable SYSDATE9 in a definition, declare it in an MVAR statement and reference it as SYSDATE9, without an ampersand, in your PROC TEMPLATE step. If you use the ampersand, the macro variable resolves when the definition is compiled instead of when ODS binds the definition to the data component.

Options

text

• is text that you can place in the definition to explain the macro variable s use. Text of this type becomes part of the compiled definition, which you can view with the SOURCE statement, whereas SAS comments do not.

NMVAR Statement

Defines a symbol that references a macro variable. ODS will convert the variable s value to a number (stored as a double) before using it. References to the macro variable are resolved when ODS binds the definition and the data component to produce an output object

Scope: You can use the NMVAR statement in the definition of a table, column, header, or footer. A macro variable that is defined in a definition is available to that definition and to all the definitions that it contains.

Featured in: Example 5 on page 539

NMVAR variable-1 < text-1 > < variable-n < text-n >>;

Required Arguments

variable

• Names a macro variable to reference in the definition. ODS will convert the variable s value to a number (stored as a double) before using it. ODS does not resolve the macro variable until it binds the definition and the data component.

• Tip: You must declare macro variables this way in a definition. For example, to use a macro variable as a number, declare it in an NMVAR statement and reference it without an ampersand. If you use the ampersand, the macro variable resolves when the definition is compiled instead of when ODS binds the definition to the data component.

Options

text

• is text that you can place in the definition to explain the macro variable s use. Text of this type becomes part of the compiled definition, which you can view with the SOURCE statement, whereas SAS comments do not.

NOTES Statement

Provides information about the column

Tip: The NOTES statement becomes part of the compiled column definition, which you can view with the SOURCE statement, whereas SAS comments do not.

NOTES text ;

Required Arguments

text

• provides information about the column.

TRANSLATE-INTO Statement

Translates the specified values to other values

TRANSLATE expression-1 INTO expression-2 < , expression-n INTO expression-m >;

Required Arguments

expression-1

• is an expression that is evaluated for each table cell in the column. It can be any expression that is valid in the WHERE statement (or the WHERE= data set option). For information on expressions that you can use in the WHERE statement, see Statements in SAS Language Reference: Dictionary . Use _VAL_ to represent the value of the current column. You may also reference symbols that you declared in a DYNAMIC, MVAR, or NVAR statement in the table definition.

  If expression-1 resolves to TRUE (a non-zero value), the translation that is specified is used for the current cell. If expression-1 is FALSE (zero), the next expression in the statement is evaluated. Thus, you can string multiple expressions together to format cells conditionally.

• Restriction: You may not reference the values of other columns in expression-1 .

• Tip: Using an expression of 1 as the last expression in the TRANSLATE “INTO statement specifies a translation for any cells that did not meet an earlier condition.

expression-2

• is an expression that specifies the value to use in the cell in place of the variable s actual value. It can be any expression that is valid in the WHERE statement (or the WHERE= data set option). For information on expressions that you can use in the WHERE statement, see Statements in SAS Language Reference: Dictionary . Use _VAL_ to represent the value of the current column. You may also reference symbols that you declared in a DYNAMIC, MVAR, or NVAR statement in the table definition.

• Restriction: expression-2 must resolve to a character value, not a numeric value.

• Restriction: You may not reference the values of other columns in expression-2 .

• Tip: When you translate a numeric value to a character value, the column definition does not try to apply the numeric format that is associated with the column. Instead, it simply writes the character value into the format field, starting at the left. If you want the value to be right-justified, use the JUSTIFY=ON attribute.

• See also: JUSTIFY= on page 384.

END Statement

Ends the definition

END ;

DEFINE FOOTER Statement

Creates a definition for a table footer

Requirement: An END statement must be the last statement in the definition.

Featured in: Example 3 on page 528 and Example 1 on page 342

See: DEFINE HEADER Statement on page 395

• DEFINE FOOTER footer-path < / STORE= libref . template-store >;

  • < footer-attribute-1 ;< footer-attribute-n ;>>

  • DYNAMIC variable-1 < text-1 > < variable-n < text-n >>;

  • MVAR variable-1 < text-1 > < variable-n < text-n >>;

  • NMVAR variable-1 < text-1 > < variable-n < text-n >>;

  • NOTES text ;

  • TEXT footer-specification ;

  • TEXT2 footer-specification ;

  • TEXT3 footer-specification ;

  • END ;

The substatements in DEFINE FOOTER and the footer attributes are the same as the substatements in DEFINE HEADER and the header attributes. For details about substatements and footer attributes, see DEFINE HEADER Statement on page 395.

DEFINE HEADER Statement

Creates a definition for a table header

Requirement: An END statement must be the last statement in the definition.

Featured in: Example 3 on page 528

• DEFINE HEADER header-path </ STORE= libref . template-store >;

  • < header-attribute-1 ;< header-attribute-n ;>>

  • DYNAMIC variable-1 < text-1 > < variable-n < text-n >>;

  • MVAR variable-1 < text-1 > < variable-n < text-n >>;

  • NMVAR variable-1 < text-1 > < variable-n < text-n >>;

  • NOTES text ;

  • TEXT header-specification ;

  • TEXT2 header-specification ;

  • TEXT3 header-specification ;

  • END ;

Required Arguments

header-path

• specifies where to store the header definition. A header-path consists of one or more names, separated by periods. Each name represents a directory in a template store. (A template store is a type of SAS file.) PROC TEMPLATE writes the definition to the first template store that you can write to in the current path.

• Restriction: If the definition is nested inside of another definition, definition-path must be a single-level name.

• Restriction: If you want to reference the definition that you are creating from another definition, then do not nest the definition inside another definition. For example, if you want to reference a header definition from multiple columns, do not define the header inside a column definition.

Options

STORE= libref . template-store

• specifies the template store in which to store the definition. If the template store does not exist, it is created.

• Restriction: If the definition is nested inside another definition, you cannot use the STORE= option for the nested definition because it is stored where the original definition is stored.

• Restriction: The STORE= option does not become part of the definition.

Header Attributes

This section lists all the attributes that you can use in a header definition. A column header spans a single column. A spanning header spans multiple columns. These two kinds of headers are defined in the same way except that a spanning header uses the START= or the END= attribute, or both.

For all attributes that support a value of ON, the following forms are equivalent:

ATTRIBUTE-NAME ATTRIBUTE-NAME=ON

For all of the attributes that support a value of variable , variable can be any variable that you declare in the table definition with the DYNAMIC, MVAR, or NMVAR statement. If the attribute is a boolean, then the value of variable should resolve to either true or false as shown in the following table:

ABBR= text

• specifies an abbreviation for the header.

• Requirement: The text must be enclosed with quotation marks.

• ODS Destinations: MARKUP

  Note: SAS includes this accessibility and compatibility feature that improves the usability of SAS for users with disabilities. This feature is related to accessibility standards for electronic information technology adopted by the U.S. Government under Section 508 of the U.S. Rehabilitation Act of 1973, as amended.

ACRONYM= text

• specifies an acronym for the header.

• Requirement: The text must be enclosed with quotation marks.

• ODS Destinations: MARKUP

  Note: SAS includes this accessibility and compatibility feature that improves the usability of SAS for users with disabilities. This feature is related to accessibility standards for electronic information technology adopted by the U.S. Government under Section 508 of the U.S. Rehabilitation Act of 1973, as amended.

ALT= text

• specifies an alternate description of the header.

• Requirement: The text must be enclosed with quotation marks.

• ODS Destinations: MARKUP

  Note: SAS includes this accessibility and compatibility feature that improves the usability of SAS for users with disabilities. This feature is related to accessibility standards for electronic information technology adopted by the U.S. Government under Section 508 of the U.S. Rehabilitation Act of 1973, as amended.

DEF_SPLIT

• specifies which special characters in headers for generic columns are to be used as split characters.

• ODS Destinations: All

END= column-name | variable

• specifies the last column that a spanning header covers.

• Default: the last column

• See also: START= on page 404

• ODS Destinations: All except OUTPUT

EXPAND= string | variable

• specifies a character to use to expand the header to fill the space over the column or columns that the header spans.

• Default: none

• Interaction: If you specify both the REPEAT=ON and EXPAND=ON attributes, then PROC TEMPLATE uses the EXPAND= attribute.

• See also: REPEAT= on page 403

  Tip: If the string or the variable that you specify contains more than one character, then PROC TEMPLATE uses only the first character.

• See also: EXPAND_PAGE= on page 401

• ODS Destinations: LISTING

EXPAND_PAGE<= ON | OFF | variable >

• specifies whether or not to expand the header to reach the sides of the page.

• Default: OFF

• See also: EXPAND= on page 400

• ODS Destinations: LISTING

FIRST_PANEL<= ON | OFF | variable >

• specifies whether or not a spanning header appears only on the first data panel if the table is too wide to fit in the space that is provided.

• Default: OFF

• Restriction: Applies only to headers, not to footers

• See also: LAST_PANEL= on page 402

• ODS Destinations: LISTING, PRINTER family, and RTF

FORCE<= ON | OFF | variable >

• specifies whether or not to try to expand the column width to accommodate the longest word in the column header. The column width can be anything between the values for the WIDTH= and WIDTH_MAX= column attributes.

• Default: ON

• See also: WIDTH= on page 406 and WIDTH_MAX= on page 388

• ODS Destinations: LISTING

GENERIC<= ON | OFF | variable >

• specifies whether or not multiple columns can use the header.

• Default: OFF

• Restriction: This attribute is primarily for writers of SAS procedures and for DATA step programmers.

• ODS Destinations: All except OUTPUT

JUST= justification | variable

• specifies the horizontal justification for the column header, where justification can be one of the following:

• LEFT

  • specifies left justification.

  • Alias: L

• RIGHT

  • specifies right justification.

  • Alias: R

• CENTER

  • specifies center justification.

  • Alias: C

• Default: The justification for the column

• ODS Destinations: All except OUTPUT

• Featured in: Example 1 on page 515

LAST_PANEL<= ON | OFF | variable >

• specifies whether or not a table footer appears only on the last data panel if the table is too wide to fit in the space that is provided.

• Default: OFF

• Restriction: Applies only to footers, not to headers

• See also: FIRST_PANEL on page 401

• ODS Destinations: LISTING, PRINTER family, and RTF

LONGDESC= string

• specifies the long description of the header.

• Requirement: The text must be enclosed with quotation marks.

• ODS Destinations: MARKUP

  Note: SAS includes this accessibility and compatibility feature that improves the usability of SAS for users with disabilities. This feature is related to accessibility standards for electronic information technology adopted by the U.S. Government under Section 508 of the U.S. Rehabilitation Act of 1973, as amended.

MAXIMIZE<=ON | OFF | variable >

• specifies whether to try to divide the text equally among all lines or to maximize the amount of text in each line when the text in the header uses more than one line. For example, if the text spans three lines, MAXIMIZE=ON might result in 45% of the text on the first line, 45% of the text on the second line, and 10% of the text on the third line. MAXIMIZE=OFF would may result in 33% of the text on each line. MAXIMIZE=ON may write lines of text that vary greatly in length. MAXIMIZE=OFF may result in using less than the full column width.

• Default: OFF

• ODS Destinations: LISTING

OVERLINE<=ON | OFF | variable >

• specifies whether or not to draw a continuous line above the header. PROC TEMPLATE uses the second formatting character to draw the line. (See the discussion of FORMCHAR= on page 417.)

• Default: OFF

• ODS Destinations: LISTING

PARENT= header-path

• specifies the header definition that the current definition inherits from. A header-path consists of one or more names, separated by periods. Each name represents a directory in a template store. (A template store is a type of SAS file.) The current definition inherits from the specified header definition in the first template store that you can read from in the current path.

  When you specify a parent, all of the attributes and statements that are specified in the parent s definition are used in the current definition unless the current definition specifically overrides them.

• ODS Destinations: All

PREFORMATTED<=ON | OFF | variable >

• specifies whether or not to treat the text as preformatted text. When text is preformatted, ODS honors line breaks as well as leading, trailing, and internal spaces. It also displays the text in a monospace font.

• Default: OFF

• Interaction: When PREFORMATTED=ON, and you are defining a table header or a footer, ODS uses the headerfixed or the footerfixed style element unless you specify another style element with the STYLE= column attribute.

  • When PREFORMATTED=ON, and you are defining a column header, ODS uses the rowheaderfixed style element unless you specify another style element with the STYLE= column attribute.

• ODS Destinations: HTML, MARKUP family, PRINTER family, and RTF

PRINT<=ON | OFF | variable >

• specifies whether or not to print the header.

• Default: ON

• Tip: When PRINT=ON, the column header becomes the label of the corresponding variable in any output data sets that the Output Destination creates if neither the column definition nor the data component provides a label.

• ODS Destinations: All

REPEAT<=ON | OFF | variable >

• specifies whether or not to repeat the text of the header until the space that is allotted for the header is filled.

• Default: OFF

• Interaction: If you specify both the REPEAT=ON and EXPAND=ON attributes, then PROC TEMPLATE uses the EXPAND= attribute.

• See also: EXPAND= on page 400

• ODS Destinations: LISTING

SPACE= positive-integer | variable

• specifies the number of blank lines to place between the current header and the next header or between the current footer and the previous footer.

• Default: 0

• Tip: A row of underlining or overlining is considered a header or a footer.

• ODS Destinations: LISTING

• Featured in: Example 3 on page 528

SPILL_ADJ<=ON | OFF | variable >

• specifies whether or not to extend the text of the header into the header space of adjacent columns.

• Default: OFF

• Interaction: FORCE=, SPILL_MARGIN=, SPILL_ADJ=, and TRUNCATE= are mutually exclusive. If you specify more than one of these attributes, then PROC TEMPLATE uses only one of these attributes. FORCE= takes precedence over the other three attributes, followed by SPILL_MARGIN=, SPILL_ADJ=, and TRUNCATE=.

• See also: FORCE= on page 401, SPILL_MARGIN= on page 403, and TRUNCATE= on page 405

• ODS Destinations: LISTING

SPILL_MARGIN<=ON | OFF | variable >

• specifies whether or not to extend the text of the header into the adjacent margin.

• Default: ON

• Restriction: SPILL_MARGIN= applies only to a spanning header that spans all the columns in a data panel.

• Interaction: FORCE=, SPILL_MARGIN=, SPILL_ADJ=, and TRUNCATE= attributes are mutually exclusive. If you specify more than one of these attributes, then PROC TEMPLATE uses only one of these attributes. FORCE= takes precedence over the other three attributes, followed by SPILL_MARGIN=, SPILL_ADJ=, and TRUNCATE=.

• See also: FORCE= on page 401, SPILL_ADJ on page 403, and TRUNCATE= on page 405

• ODS Destinations: LISTING

SPLIT= character | variable

• specifies the split character for the header. PROC TEMPLATE starts a new line when it reaches that character and continues the header on the next line. The split character itself is not part of the header although each occurrence of the split character counts toward the maximum length for a label.

• Tip: The first character in a header is automatically treated as a split character if it is not one of the following:

  • an alphanumeric character

  • a blank

  • an underscore (_)

  • a hyphen (-).

• ODS Destinations: All except OUTPUT

START= column-name | variable

• specifies the first column that a spanning header covers.

• Default: the first column

• See also: END= on page 400

• ODS Destinations: All except OUTPUT

STYLE=< style-element-name ><[ style-attribute-specification(s) ]>

• specifies the style element and any changes to its attributes to use for the header.

• Requirement: The STYLE= option requires either a style-attribute-specification or a style-element-name .

• Tip: You can use braces ({ and }) instead of square brackets ([ and ]).

• Tip: If you use the STYLE= attribute inside a quoted string, then you must add a space before or after the carriage return to prevent errors. SAS does not interpret a carriage return as a space. You must explicitly specify spaces in your quoted strings.

  style-element-name

  • is the name of the style element to use to produce the header. The style element must be part of a style definition that is registered with the Output Delivery System. SAS provides some style definitions. You can create your own style definitions by using PROC TEMPLATE (see DEFINE STYLE Statement on page 288). By default, ODS produces different parts of ODS output with different elements. For example, by default, a table header is displayed with the style element header . The style elements that you would be most likely to use with the STYLE= attribute for a table header are as follows :

    • header

    • headerfixed

    • headerempty

    • headeremphasis

    • headeremphasisfixed

    • headerstrong

    • headerstrongfixed

  • The style elements that you would be most likely to use with the STYLE= attribute for a table footer are as follows:

    • footer

    • footerfixed

    • footerempty

    • footeremphasis

    • footeremphasisfixed

    • footerstrong

    • footerstrongfixed

  • The style elements that you would be most likely to use with the STYLE= attribute for a column header are as follows:

    • rowheader

    • rowheaderfixed

    • rowheaderempty

    • rowheaderemphasis

    • rowheaderemphasisfixed

    • rowheaderstrong

    • rowheaderstrongfixed

  • The style element provides the basis for displaying the header. Additional style attributes that you provide can modify the display.

  • style-element-name can be either the name of a style element or a variable whose value is a style element.

  • Default: header

  • See also: Viewing the Contents of a Style Definition on page 319.

  • See also: The Default Style Definition for HTML and Markup Languages on page 320.

• style-attribute-specification

  • describes the style attribute to change. Each style-attribute-specification has this general form:

    • style-attribute-name = style-attribute-value

  • ODS destinations: HTML, PRINTER family, and RTF

  • Featured in: Example 1 on page 515 and Example 3 on page 528

  • See also: Style Attributes and Their Values on page 292

TRUNCATE<=ON | OFF | variable >

• specifies whether or not to start a new header line in the middle of a word.

• ON

  • starts a new line of the header when the text fills the specified column width.

• OFF

  • extends the width of the column to accommodate the longest word in the column header, if possible.

    Note: TRUNCATE=OFF is the same as FORCE=ON.

• Default: OFF

• Interaction: If you specify FORCE=, SPILL_MARGIN=, or SPILL_ADJ=, then the TRUNCATE= attribute is ignored.

• See also: FORCE= on page 401, SPILL_MARGIN= on page 403, and SPILL_ADJ= on page 403

• ODS Destinations: LISTING

UNDERLINE<=ON | OFF | variable >

• specifies whether or not to draw a continuous line below the header. PROC TEMPLATE uses the second formatting character to draw the line.

• Default: OFF

• Main discussion: See the discussion of FORMCHAR= on page 417.

• ODS Destinations: LISTING

VJUST= justification | variable

• Specifies vertical justification for the header. justification can be one of the following:

• TOP

  • places the header as high as possible.

  • Alias: T

• CENTER

  • centers the header vertically.

  • Alias: C

• BOTTOM

  • places the header as low as possible.

  • Alias: B

• Default: BOTTOM

• ODS Destinations: HTML and PRINTER family

WIDTH= positive-integer | variable

• specifies the width of the header in characters.

• Default: If you do not specify a width, PROC TEMPLATE uses the column width.

• Tip: If you want a vertical header, specify a width of 1.

• ODS Destinations: LISTING

DYNAMIC Statement

Defines a symbol that references a value that the data component supplies from the procedure or DATA step

Scope: You can use the DYNAMIC statement in the definition of a table, column, header, or footer. A dynamic variable that is defined in a definition is available to that definition and to all the definitions that it contains.

Featured in: Example 1 on page 342 and Example 2 on page 348

DYNAMIC variable-1 < text-1 > < variable-n < text-n >>;

Required Arguments

variable

• Names a variable that the data component supplies. ODS resolves the value of the variable when it binds the definition and the data component.

• Tip: Dynamic variables are most useful to the authors of SAS procedures and to DATA step programmers.

Options

text

• is text that you can place in the definition to explain the dynamic variable s use. Text of this type becomes part of the compiled definition, which you can view with the SOURCE statement, whereas SAS comments do not.

MVAR Statement

Defines a symbol that references a macro variable. ODS will use the value of the variable as a string. References to the macro variable are resolved when ODS binds the definition and the data component to produce an output object

Scope: You can use the MVAR statement in the definition of a table, column, header, or footer. A macro variable that is defined in a definition is available to that definition and to all the definitions that it contains.

Featured in: Example 3 on page 528 and Example 1 on page 342

MVAR variable-1 < text-1 > < variable-n < text-n >>;

Required Arguments

variable

• Names a macro variable to reference in the definition. ODS will use the value of the macro variable as a string. ODS does not resolve the value of the macro variable until it binds the definition and the data component.

• Tip: You must declare macro variables this way in a definition. For example, to use the automatic macro variable SYSDATE9 in a definition, declare it in an MVAR statement and reference it as SYSDATE9, without an ampersand, in your PROC TEMPLATE step. If you use the ampersand, the macro variable resolves when the definition is compiled instead of when ODS binds the definition to the data component.

Options

text

• is text that you can place in the definition to explain the macro variable s use. Text of this type becomes part of the compiled definition, which you can view with the SOURCE statement, whereas SAS comments do not.

NMVAR Statement

Defines a symbol that references a macro variable. ODS will convert the variable s value to a number (stored as a double) before using it. References to the macro variable are resolved when ODS binds the definition and the data component to produce an output object

Scope: You can use the NMVAR statement in the definition of a table, column, header, or footer. A macro variable that is defined in a definition is available to that definition and to all the definitions that it contains.

Featured in: Example 5 on page 539

NMVAR variable-1 < text-1 > < variable-n < text-n >>;

Required Arguments

variable

• Names a macro variable to reference in the definition. ODS will convert the variable s value to a number (stored as a double) before using it. ODS does not resolve the macro variable until it binds the definition and the data component.

• Tip: You must declare macro variables this way in a definition. For example, to use a macro variable as a number, declare it in an NMVAR statement and reference it without an ampersand. If you use the ampersand, the macro variable resolves when the definition is compiled instead of when ODS binds the definition to the data component.

Options

text

• is text that you can place in the definition to explain the macro variable s use. Text of this type becomes part of the compiled definition, which you can view with the SOURCE statement, whereas SAS comments do not.

NOTES Statement

Provides information about the header

Tip: The NOTES statement becomes part of the compiled header definition, which you can view with the SOURCE statement, whereas SAS comments do not.

NOTES text ;

Required Arguments

text

• provides information about the header.

TEXT Statement

Specifies the text of the header or the label of a variable in an output data set

Featured in: Example 3 on page 528

TEXT header-specification(s) ;

Required Arguments

header-specification(s)

• specifies the text of the header. Each header-specification can be one of the following:

• _LABEL_

  • uses the label of the object that the header applies to as the text of the header. For example, if the header is for a column, _LABEL_ specifies the label for the variable that is associated with the column. If the header is for a table, _LABEL_ specifies the label for the data set that is associated with the table.

• text-specification(s)

  • specifies the text to use in the header. Each text-specification can be one of the following:

    • a quoted string

    • a variable, followed by an optional format. The variable can be any variable that is declared in a DYNAMIC, MVAR, or NMVAR statement.

  • Note: If the first character in a quoted string is neither a blank character nor an alphanumeric character, and SPLIT is not in effect, the TEXT statement treats that character as the split character. (See the discussion of SPLIT= on page 404.)

  • Default: If you do not use a TEXT statement, the text of the header is the label of the object that the header applies to.

  • Tip: If the quoted string is a blank and it is the only item in the header specification, the header is a blank line.

  • Featured in: Example 3 on page 528

TEXT2 Statement

Provides an alternative header to use in the Listing output if the header that is provided by the TEXT statement is too long

See: TEXT Statement on page 409

TEXT3 Statement

Provides an alternative header to use in the Listing output if the header that is provided by the TEXT2 statement is too long

See: TEXT Statement on page 409

DEFINE TABLE Statement

Creates a table definition

Requirement: An END statement must be the last statement in the definition.

Interaction: A table definition can contain one or more column, header, or footer definitions.

Featured in: Example 3 on page 528 and Example 5 on page 539

• DEFINE TABLE table-path </ STORE= libref . template-store >;

  • < table-attribute-1 ;< table-attribute-n ;>>

  • CELLSTYLE expression-1 AS < style-element-name ><[ style-attribute-specification(s) ] >< , expression-n AS < style-element-name ><[ style-attribute-specification(s) ]>>;

  • COLUMN column(s) ;

  • DEFINE definition-type definition-name </ option(s) >; statements-and-attributes

    • END ;

  • DYNAMIC variable-1 < text-1 > < variable-n < text-n >>;

  • FOOTER footer-name(s) ;

  • HEADER header-name(s) ;

  • MVAR variable-1 < text-1 > < variable-n < text-n >>;

  • NMVAR variable-1 < text-1 > < variable-n < text-n >>;

  • NOTES text ;

  • TRANSLATE expression-1 INTO expression-2 < , expression-n INTO expression-m ;>

  • END ;

Required Arguments

table-path

• specifies where to store the table definition. A table-path consists of one or more names, separated by periods. Each name represents a directory in a template store. (A template store is a type of SAS file.) PROC TEMPLATE writes the definition to the first template store in the current path that you can write to.

Options

STORE= libref . template-store

• specifies the template store in which to store the definition. If the template store does not exist, it is created.

• Restriction: The STORE= option does not become part of the definition.

Table Attributes

This section lists all the attributes that you can use in a table definition. For all attributes that support a value of ON, the following forms are equivalent:

ATTRIBUTE-NAME ATTRIBUTE-NAME=ON

For all of the attributes that support a value of variable , variable can be any variable that you declare in the table definition with the DYNAMIC, MVAR, or NMVAR statement. If the attribute is a boolean, then the value of variable should resolve to either true or false as shown in the following table:

ALT= text

• specifies an alternate description of the table.

• Requirement: The text must be enclosed with quotation marks.

• ODS Destinations: MARKUP

  Note: SAS includes this accessibility and compatibility feature that improves the usability of SAS for users with disabilities. This feature is related to accessibility standards for electronic information technology adopted by the U.S. Government under Section 508 of the U.S. Rehabilitation Act of 1973, as amended.

BALANCE <=ON | OFF | variable >

• specifies whether or not to try to place the same number of columns in each data panel if the entire table does not fit in one data panel.

• Default: OFF

• ODS Destinations: LISTING, PRINTER family, and RTF

BYLINE <=ON | OFF | variable >

• specifies whether or not to print the current byline before the table.

• Default: OFF

• Restriction: This attributes applies only if the table is not the first one on the page. If BY- group processing is in effect, a byline automatically precedes the first table on the page.

• ODS Destinations: All except OUTPUT

CENTER <=ON | OFF | variable >

• specifies whether or not to center each data panel independently if the entire table does not fit in the space that is provided.

• Default: ON

• ODS Destinations: LISTING, PRINTER family, and RTF

CLASSLEVELS <=ON | OFF | variable >

• specifies whether or not to suppress blanking the value in a column that is marked with the BLANK_DUPS column attribute if the value changes in a previous column that is also marked with the BLANK_DUPS attribute.

• Default: OFF

• ODS Destinations: All except OUTPUT

• Featured in: Example 1 on page 342

COL_SPACE_MAX= positive-integer | variable

• specifies the maximum number of blank characters to place between the columns.

• Default: 4

• ODS Destinations: LISTING

COL_SPACE_MIN= positive-integer | variable

• specifies the minimum number of blank characters to place between the columns.

• Default: 2

• ODS Destinations: LISTING

CONTENTS <=ON | OFF | variable >

• specifies whether or not to place the output object in a table of contents, if you create a table of contents.

• Default: ON

• ODS Destinations: HTML, PDF, and PRINTER/PS PDFMARK

CONTENTS_LABEL= string | variable

• specifies the label to use for the output object in the contents file, the Results window, and the trace record.

• Default: If the SAS system option LABEL is in effect, the default label is the object s label. If LABEL is not in effect, the default label is the object s name.

• ODS Destinations: HTML, PDF, and PRINTER/PS PDFMARK

CONTROL= column-name | variable

• specifies the name of the column whose values provide formatting information about the space before each row of the definition. The value of CONTROL= should be the name of a column of type character with a length equal to 1.

  Column Control Value	Result
  a digit from 1-9	the specified number of blank lines precedes the current row
  a hyphen (-)	a row of underlining precedes the current row
  b or B	ODS tries to insert a panel break if the entire table does not fit in the space that is provided

  Default: none

  ODS Destinations: All except OUTPUT

DATA_FORMAT_OVERRIDE<=ON | OFF | variable >

• specifies which format to use if both a column definition and a data component specify a format.

  ON

  • uses the format that the data component specifies.

• OFF

  • use the format that the column definition specifies.

• Default: OFF

  ODS Destinations: All

DOUBLE_SPACE<=ON | OFF | variable >

• specifies whether or not to double space between the rows of the table.

  Default: OFF

  ODS Destinations: LISTING

  Featured in: Example 1 on page 515 and Example 3 on page 528

EVEN<=ON | OFF | variable >

• specifies whether or not extra space is evenly divided among all columns of the table.

  Default: OFF

  ODS Destinations: LISTING

FOOTER_SPACE=0 | 1 | 2 | variable

• specifies the number of blank lines to place between the last row of data and the first row of the table footer.

  Default: 1

  ODS Destinations: LISTING

FORMCHAR= string | variable

• defines the characters to use as the line-drawing characters in the table. Currently, PROC TEMPLATE uses only the second of the 20 possible formatting characters. This formatting character is used for underlining and overlining. To change the second formatting character, you must specify both the first and second formatting characters. For example, the following option assigns the asterisk (*) to the first formatting character, the plus sign (+) to the second character, and does not alter the remaining characters:

  formchar=*+

  Default: The SAS system option FORMCHAR= specifies the default formatting characters.

  Tip: You can use any character in formatting-characters, including hexadecimal characters. If you use hexadecimal characters, then you must put an x after the closing quote. For example, the following option assigns the hexadecimal character 2D to the first formatting character, the hexadecimal character 7C to the second character, and does not alter the remaining characters:

  formchar= 2D7C x

• ODS Destinations: LISTING

HEADER_SPACE=0 | 1 | 2 | variable

• specifies the number of blank lines to place between the last row of headers and the first row of data. A row of underscores is a header.

• Default: 1

• ODS Destinations: LISTING

JUSTIFY<=ON | OFF | variable >

• specifies whether to justify the format fields within the columns or to justify the values within the columns without regard to the format fields.

• Default: OFF

• Interaction: JUSTIFY=ON can interfere with decimal alignment.

• Interaction: If the column is numeric, then values are aligned to the right if you specify JUSTIFY=OFF and JUST=C.

• Interaction: All destinations except LISTING justify the values in columns as if JUSTIFY=ON for JUST=R and JUST=L.

• Tip: If you translate numeric data to character data, you might need to use JUSTIFY= to align the data.

• Main discussion: How Are Values in Table Columns Justified? on page 512

• ODS Destinations: LISTING

LABEL= text | variable

• specifies a label for the table.

• Default: PROC TEMPLATE uses the first of the following that it finds:

  • a label that the table definition provides

  • a label that the data component provides

  • the first spanning header in the table.

• ODS Destinations: All

LONGDESC= string

• specifies the long description of the table.

• Requirement: The text must be enclosed with quotation marks.

• ODS Destinations: MARKUP

  Note: SAS includes this accessibility and compatibility feature that improves the usability of SAS for users with disabilities. This feature is related to accessibility standards for electronic information technology adopted by the U.S. Government under Section 508 of the U.S. Rehabilitation Act of 1973, as amended.

NEWPAGE<=ON | OFF | variable >

• specifies whether or not to force a new page before printing the table.

• Default: OFF

• Restriction: If the table is the first item on the page, ODS ignores this attribute.

• ODS Destinations: All except OUTPUT

ORDER_DATA<=ON | OFF | variable >

• specifies whether or not to order the columns by their order in the data component.

• Default: OFF

  • When ORDER_DATA=OFF, the default order for columns is the order that they are specified in the COLUMN statement. If you do not use a COLUMN statement, the default order for columns is the order in which you define them in the definition.

• Tip: The Output destination always uses the order of the columns in the data component when it creates an output data set.

• Interaction: ORDER_DATA is most useful for ordering generic columns.

• ODS Destinations: All except OUTPUT

OVERLINE<=ON | OFF | variable >

• specifies whether or not to draw a continuous line above the first table footer or, if there is no table footer, below the last row of data on a page. PROC TEMPLATE uses the second formatting character to draw the line.

• Default: OFF

• Main discussion: See the discussion of FORMCHAR= on page 417.

• See also: UNDERLINE= on page 421 (for tables), UNDERLINE= on page 387 (for columns), and OVERLINE= on page 385 (for columns)

• ODS Destinations: LISTING

• Featured in: Example 1 on page 515

PANELS= positive-integer | variable

• specifies the number of sets of columns to place on a page. If the width of all the columns is less than half of the linesize, you can display the data in multiple sets of columns so that rows that would otherwise appear on multiple pages appear on the same page.

• Tip: If the number of panels that is specified is larger than the number of panels that can fit on the page, the definition creates as many panels as it can. Let the table definition put your data in the maximum number of panels that can fit on the page by specifying a large number of panels (for example, 99).

• ODS Destinations: LISTING and PRINTER family

PANEL_SPACE= positive-integer | variable

• specifies the number of blank characters to place between sets of columns when PANELS= is in effect.

• Default: 2

• ODS Destinations: LISTING

PARENT= table-path

• specifies the table that the current definition inherits from. A table-path consists of one or more names, separated by periods. Each name represents a directory in a template store. (A template store is a type of SAS file.) The current definition inherits from the specified table in the first template store in the current path that you can read from.

  When you specify a parent, all of the attributes and statements that are specified in the parent s definition are used in the current definition unless the current definition overrides them.

• ODS Destinations: All

PRINT_FOOTERS<=ON | OFF | variable >

• specifies whether or not to print table footers and any overlining of the table footers.

• Default: ON

• See also: OVERLINE= on page 419

• ODS Destinations: All except OUTPUT

PRINT_HEADERS<=ON | OFF | variable >

• specifies whether or not to print the table headers and any underlining of the table headers.

• Default: ON

• Interaction: When used in a table definition, PRINT_HEADERS affects only headers for the table, not the headers for individual columns. (See the discussion of the PRINT_HEADERS column attribute on page 386.)

• See also: UNDERLINE= on page 421

• ODS Destinations: All except OUTPUT

REQUIRED_SPACE= positive-integer | variable

• specifies the number of lines that must be available on the page in order to print the body of the table (The body of the table is the part of the table that contains the data. It does not include headers and footers.)

• Default: 3

• ODS Destinations: LISTING and PRINTER family

SPLIT_STACK<=ON | OFF | variable >

• specifies whether or not to split a long stacked column across page boundaries.

• Default: OFF

• ODS Destinations: LISTING

STYLE=< style-element-name ><[ style-attribute-specification(s) ]>

• specifies the style element and any changes to its attributes to use for the table.

• style-element-name

  • is the name of the style element to use to display the table. The style element must be part of a style definition that is registered with the Output Delivery System. SAS provides some style definitions. You can create your own style definitions with PROC TEMPLATE (see DEFINE STYLE Statement on page 288). By default, ODS produces different parts of ODS output with different elements. For example, by default, a table is produced with the style element table . The style definitions that SAS provides do not provide another style element that you would be likely to want to use instead of table . However, you may have a user -defined style element at your site that would be appropriate to specify.

    The style element provides the basis for displaying the table. Additional style attributes that you provide can modify the display.

    style-element-name can be either the name of a style element or a variable whose value is a style element.

  • See also: Viewing the Contents of a Style Definition on page 319

  • See also: The Default Style Definition for HTML and Markup Languages on page 320.

• style-attribute-specification

  • describes the style attribute to change. Each style-attribute-specification has this general form:

    • style-attribute-name = style-attribute-value

  • See also: Style Attributes and Their Values on page 292.

• Default: table

• Requirement: With the STYLE= option, you must specify either a style-attribute-specification or a style-element-name .

• Tip: You can use braces ({ and }) instead of square brackets ([ and ]).

• Tip: If you use the STYLE= attribute inside a quoted string, then you must add a space before or after the carriage return to prevent errors. SAS does not interpret a carriage return as a space. You must explicitly specify spaces in your quoted strings.

• ODS Destinations: HTML, MARKUP family, PRINTER family, and RTF

TOP_SPACE= positive-integer | variable

• specifies the number of lines to place between the previous output object and the current one.

• Default: 1

• ODS Destinations: LISTING and PRINTER family

TYPE= string | variable

• specifies special type of SAS data set.

• Restriction: PROC TEMPLATE does not verify the following:

  • a SAS data set type that you specify is a valid data set type

  • the structure of the data set that you create is appropriate for the type that you have assigned.

• Tip: Most SAS data sets have no special type. However, certain SAS procedures, like the CORR procedure, can create a number of special SAS data sets. In addition, SAS/STAT software and SAS/EIS software support special data set types.

• ODS Destination: OUTPUT

UNDERLINE<=ON | OFF | variable >

• specifies whether or not to draw a continuous line under the last table header (or, if there is no table header, then above the first row of data on a page). PROC TEMPLATE uses the second formatting character to draw the line.

• Default: OFF

• Main discussion: See the discussion of FORMCHAR= on page 417.

• See also: OVERLINE= on page 419 (for tables), UNDERLINE= on page 387 (for columns), and OVERLINE= on page 385 (for columns)

• ODS Destinations: LISTING

• Featured in: Example 1 on page 515 and Example 3 on page 528

USE_FORMAT_DEFAULTS<=ON | OFF | variable >

• specifies the source of the values for the format width and the decimal width if they are not specified.

• ON

  • uses the default values, if any, that are associated with the format name.

• OFF

  • uses the PROC TEMPLATE defaults.

• Default: OFF

• ODS Destinations: All except OUTPUT

USE_NAME<=ON | OFF | variable >

• uses the column name as the column header if neither the column definition nor the data component specifies a header.

• Default: OFF

• Tip: Use this attribute when column names are derived from a data set and the columns are generic.

• ODS Destinations: All except OUTPUT

WRAP<=ON | OFF | variable >

• specifies whether to split a wide table into multiple data panels, or to wrap each row of the table so that an entire row is printed before the next row starts.

• Default: OFF

• Interaction: When ODS wraps the rows of a table, it does not place multiple values in any column that contains an ID column.

• See also: WRAP_SPACE= on page 422 and ID= on page 383

• ODS Destinations: LISTING and PRINTER family

WRAP_SPACE<=ON | OFF | variable >

• specifies whether or not to double space after the last line of a single row of the table when the row is wrapped onto more than one line.

• Default: OFF

• See also: WRAP= on page 422

• ODS Destinations: LISTING, PRINTER family, and RTF

CELLSTYLE “AS Statement

Sets the style element of the cells in the table according to the values of the variables. Use this statement to set the presentation characteristics (such as foreground color, font face, flyover) of individual cells

Featured in: Example 5 on page 539

• CELLSTYLE expression-1 AS < style-element-name ><[ style-attribute-specification(s) ]> < , expression-n AS < style-element-name ><[ style-attribute-specification(s) ]>>;

Required Arguments

expression

• is an expression that is evaluated for each table cell that contains a variable. It can be any expression that is valid in the WHERE statement (or the WHERE= data set option). For information on expressions that you can use in the WHERE statement, see Statements in SAS Language Reference: Dictionary . Use _VAL_ to represent the value of the current column. You may also reference symbols that you declared in a DYNAMIC, MVAR, or NVAR statement in the table definition.

  If expression resolves to TRUE (a non-zero value), the style element that is specified is used for the current cell. If expression is FALSE (zero), the next expression in the statement is evaluated. Thus, you can string multiple expressions together to format cells conditionally.

• Restriction: You can not reference the values of other columns in expression .

• Tip: Using an expression of 1 as the last expression in the CELLSTYLE-AS statement sets the style element for any cells that did not meet an earlier condition.

style-attribute-specification

• describes a style attribute to set. Each style-attribute-specification has this general form:

  • style-attribute-name = style-attribute-value

• For information on the style attributes that you can set in a table definition, see Style Attributes and Their Values on page 292.

Options

style-element-name

• is the name of a style element that is part of a style definition that is registered with the Output Delivery System. SAS provides some style definitions. You can create your own style definitions and style elements with PROC TEMPLATE. (See DEFINE STYLE Statement on page 288.)

  The style elements that you would be most likely to use with the CELLSTYLE-AS statement are

  • data

  • datafixed

  • dataempty

  • dataemphasis

  • dataemphasisfixed

  • datastrong

  • datastrongfixed.

• The style element provides the basis for displaying the cell. Additional style attributes that you provide can modify the display.

COLUMN Statement

Declares a symbol as a column in the table and specifies the order of the columns

Featured in: Example 3 on page 528

COLUMN column(s) ;

Required Arguments

column

• is one or more columns. If the column is defined outside the current table definition, you must reference it by its path in the template store. Columns in the definition are laid out from left to right in the same order that you specify them in the COLUMN statement.

• Default: If you do not use a COLUMN statement, ODS makes a column for each column definition (DEFINE COLUMN statement), and places the columns in the same order that the column definitions have in the table definition.

  If you use a COLUMN statement but do not use a DEFINE COLUMN statement for any of the columns, ODS uses a default column definition that is based on the type of data in the column.

• Interaction: If you specify the column attribute PRINT=OFF, then you turn off the value of a column if it is part of a stacked column. If all columns in a stacked column have PRINT=OFF set, then the entire column is removed from the table.

• Tip: You can use a list of variable names, such as DAY1 “DAY10, to specify multiple variables.

• Main discussion: Stacking Values for Two or More Variables

  • To stack values for two or more variables in the same column, put parentheses around the variables that you want to stack. In such a case, the column header for the first column inside the parentheses becomes the header for the column that contains all the variables inside parentheses. For example, the following COLUMN statement produces a definition in which

    • the value of NAME is in the first column by itself.

    • the values of CITY and STATE appear in the second column with CITY above STATE. The header for this column is the header that is associated with CITY.

    • the values HOMEPHONE and WORKPHONE appear in the third column with HOMEPHONE above WORKPHONE. The header for this column is the header that is associated with HOMEPHONE.

      column name (city state) (homephone workphone);

  • You can use the asterisk (*) in the COLUMN statement to change the layout of stacking variables. An asterisk between groups of variables in parentheses stacks the first item in the first set of parentheses above the first item in the next set of parentheses, and so on until the last group of parentheses is reached. Then, the second item in the first group is stacked above the second item in the second group, and so on. For example, the following COLUMN statement produces a report in which

    • the value of NAME is in the first column by itself.

    • the values of CITY and HOMEPHONE appear in the second column with CITY above HOMEPHONE. The header for this column is the header that is associated with CITY.

    • the values STATE and WORKPHONE appear in the third column with STATE above WORKPHONE. The header for this column is the header that is associated with STATE.

      column name (city state) * (homephone workphone);

DEFINE Statement

Creates a definition inside a table definition

Main discussion: DEFINE COLUMN Statement on page 374, DEFINE FOOTER Statement on page 395, and DEFINE HEADER Statement on page 395

DEFINE definition-type definition-name </ option(s) >; statements-and-attributes

• END ;

Required Arguments

definition-type

• specifies the type of definition to create, where definition-type is one of the following:

• COLUMN

• FOOTER

• HEADER

• The definition-type determines what other statements and what attributes can go in the definition. For details, see the documentation for the corresponding DEFINE statement.

definition-name

• specifies the name of the new object.

• Restriction: definition-name must be a single-level name.

  • Note: If you want to reference the definition that you are creating from another definition, you must create it outside the table definition.

Options

NOLIST

• preserves the definition-type when inheriting it from another table definition.

• Tip: If you specify an existing definition-name without using the NOLIST option, then the definition is overwritten.

DYNAMIC Statement

Defines a symbol that references a value that the data component supplies from the procedure or DATA step

Scope: You can use the DYNAMIC statement in the definition of a table, column, header, or footer. A dynamic variable that is defined in a definition is available to that definition and to all the definitions that it contains.

Featured in: Example 1 on page 342 and Example 2 on page 348

DYNAMIC variable-1 < text-1 > < variable-n < text-n >>;

Required Arguments

variable

• Names a variable that the data component supplies. ODS resolves the value of the variable when it binds the definition and the data component.

• Tip: Dynamic variables are most useful to the authors of SAS procedures and to DATA step programmers.

Options

text

• is text that you can place in the definition to explain the dynamic variable s use. Text of this type becomes part of the compiled definition, which you can view with the SOURCE statement, whereas SAS comments do not.

FOOTER Statement

Declares a symbol as a footer in the table and specifies the order of the footers

FOOTER footer-specification(s) ;

Required Arguments

footer-specification

• is one or more footers. If the footer is defined outside the current table definition, you must reference it by its path in the template store. Footers in the definition are laid out from top to bottom in the same order that you specify them in the FOOTER statement. Each footer-specification can be

• " string "

  • specifies the text to use for the footer. If you use a string, you do not need to use a DEFINE FOOTER statement. However, you cannot specify any footer attributes except for a split character. If the SPLIT= attribute is not in effect and if the first character of the footer that you specify is neither a blank character nor an alphanumeric character, PROC TEMPLATE treats it as the split character.

  • See also: SPLIT=.

• footer-path

  • is the path of the footer definition to use. A footer-path consists of one or more names, separated by periods. Each name represents a directory in a template store. (A template store is a type of SAS file.)

• _LABEL_

  • uses the label of the output object as the footer. Each SAS procedure specifies a label for each output object that it creates. The DATA step uses the value of the OBJECTLABEL= option as the label of the output object. If OBJECTLABEL= is not specified, it uses the text of the first TITLE statement as the label.

• Default: If you do not use a FOOTER statement, ODS makes a footer for each footer definition (DEFINE FOOTER statement), and places the footers in the same order that the footer definitions have in the table definition.

HEADER Statement

Declares a symbol as a header in the table and specifies the order of the headers

HEADER header-specification(s) ;

Required Arguments

header-specification

• is one or more headers. If the header is defined outside the current table definition, you must reference it by its path in the template store. Headers in the definition are laid out from top to bottom in the same order that you specify them in the HEADER statement. Each header-specification can be

• " string "

  • specifies the text to use for the header. If you use a string, you do not need to use a DEFINE HEADER statement. However, you cannot specify any header attributes except for a split character. If the SPLIT= header attribute is not in effect and if the first character of the header that you specify is neither a blank character nor an alphanumeric character, PROC TEMPLATE treats it as the split character.

    See also: SPLIT=.

• header-path

  • is the path of the header definition to use. A header-path consists of one or more names, separated by periods. Each name represents a directory in a template store. (A template store is a type of SAS file.)

• _LABEL_

  • uses the label of the output object as the header. Each SAS procedure specifies a label for each output object that it creates. The DATA step uses the value of the OBJECTLABEL= option as the label of the output object. If OBJECTLABEL= is not specified, it uses the text of the first TITLE statement as the label.

• Default: If you do not use a HEADER statement, then ODS makes a header for each header definition (DEFINE HEADER statement), and places the headers in the same order that the header definitions have in the table definition.

• Featured in: Example 3 on page 528

MVAR Statement

Defines a symbol that references a macro variable. ODS will use the value of the variable as a string. References to the macro variable are resolved when ODS binds the definition and the data component to produce an output object

Scope: You can use the MVAR statement in the definition of a table, column, header, or footer. A macro variable that is defined in a definition is available to that definition and to all the definitions that it contains.

Featured in: Example 3 on page 528 and Example 1 on page 342

MVAR variable-1 < text-1 > < variable-n < text-n >>;

Required Arguments

variable

• Names a macro variable to reference in the definition. ODS will use the value of the macro variable as a string. ODS does not resolve the value of the macro variable until it binds the definition and the data component.

• Tip: You must declare macro variables this way in a definition. For example, to use the automatic macro variable SYSDATE9 in a definition, declare it in an MVAR statement and reference it as SYSDATE9, without an ampersand, in your PROC TEMPLATE step. If you use the ampersand, the macro variable resolves when the definition is compiled instead of when ODS binds the definition to the data component.

Options

text

• is text that you can place in the definition to explain the macro variable s use. Text of this type becomes part of the compiled definition, which you can view with the SOURCE statement, whereas SAS comments do not.

NMVAR Statement

Defines a symbol that references a macro variable. ODS will convert the variable s value to a number (stored as a double) before using it. References to the macro variable are resolved when ODS binds the definition and the data component to produce an output object

Scope: You can use the NMVAR statement in the definition of a table, column, header, or footer. A macro variable that is defined in a definition is available to that definition and to all the definitions that it contains.

Featured in: Example 5 on page 539

NMVAR variable-1 < text-1 > < variable-n < text-n >>;

Required Arguments

variable

• Names a macro variable to reference in the definition. ODS will convert the variable s value to a number (stored as a double) before using it. ODS does not resolve the macro variable until it binds the definition and the data component.

• Tip: You must declare macro variables this way in a definition. For example, to use a macro variable as a number, declare it in an NMVAR statement and reference it without an ampersand. If you use the ampersand, the macro variable resolves when the definition is compiled instead of when ODS binds the definition to the data component.

Options

text

• is text that you can place in the definition to explain the macro variable s use. Text of this type becomes part of the compiled definition, which you can view with the SOURCE statement, whereas SAS comments do not.

NOTES Statement

Provides information about the table

Tip: The NOTES statement becomes part of the compiled column definition, which you can view with the SOURCE statement, whereas SAS comments do not.

Featured in: Example 5 on page 539

NOTES text ;

Required Arguments

text

• provides information about the table.

TRANSLATE-INTO Statement

Translates the specified numeric values to other values

Restriction: The TRANSLATE-INTO statement in a table definition applies only to numeric variables. To translate the values of a character variable, use TRANSLATE-INTO in the definition of that column. (See DEFINE COLUMN Statement on page 374).

Featured in: Example 5 on page 539

TRANSLATE expression-1 INTO expression-2 < , expression-n INTO expression-m >;

Required Arguments

expression-1

• is an expression that is evaluated for each table cell that contains a numeric variable. It can be any numeric expression that is valid in the WHERE statement (or the WHERE= data set option). For information on expressions that you can use in the WHERE statement, see Statements in SAS Language Reference: Dictionary . Use _VAL_ to represent the value of the current column. You may also reference symbols that you declared in a DYNAMIC, MVAR, or NVAR statement in the table definition.

  If expression-1 resolves to TRUE (a non-zero value), the translation that is specified is used for the current cell. If expression-1 is FALSE (zero), the next expression in the statement is evaluated. Thus, you can string multiple expressions together to format cells conditionally.

  Restriction: You may not reference the values of other columns in expression-1 .

  Tip: Using an expression of 1 as the last expression in the TRANSLATE “INTO statement specifies a translation for any cells that did not meet an earlier condition.

expression-2

• is an expression that specifies the value to use in the cell in place of the variable s actual value. It can be any expression that is valid in the WHERE statement (or the WHERE= data set option). For information on expressions that you can use in the WHERE statement, see Statements in SAS Language Reference: Dictionary . Use _VAL_ to represent the value of the current column. You may also reference symbols that you declared in a DYNAMIC, MVAR, or NVAR statement in the table definition.

• Restriction: expression-2 must resolve to a character value, not a numeric value.

• Restriction: You may not reference the values of other columns in expression-2 .

• Tip: When you translate a numeric value to a character value, the table definition does not try to apply the numeric format that is associated with the column. Instead, it simply writes the character value into the formatted field, starting at the left. If you want the value to be right-justified, use the JUSTIFY=ON attribute.

• See also: JUSTIFY= on page 418.

END Statement

Ends the table definition

END ;

ODS Output Object Table Names

• ODS Table Names and the Base SAS Procedures that Produce Them on page 430

• ODS Table Names and the SAS/STAT Procedures that Produce Them on page 437

• ODS Table Names and the SAS/ETS Procedures that Produce Them on page 490

ODS Table Names and the Base SAS Procedures that Produce Them

The following table lists the output object table names which Base SAS procedures produce. The table provides the name of each table, a description of what the table contains, and the option, if any, that creates the output object table. For more information about Base SAS procedures, see Base SAS Procedures Guide .

ODS Table Names and the SAS/STAT Procedures that Produce Them

The following table lists the output object table names which SAS/STAT procedures produce. You must license SAS/STAT software in order to produce these output objects. The table provides the name of each table, a description of what the table contains, and the option, if any, that creates the output object table. For information about SAS/STAT procedures, see