Introduction to SAS/Graph

ANNOTATE= data sets

ANNOTATE= data sets contain the commands or functions that instruct SAS/GRAPH software on how to enhance your output. Two basic functions allow for moving and drawing. Other functions are used to position labels. More complex functions are used to create bars, pies, and polygons. An ANNOTATE= data set consists of one observation per function, and predetermined variable names are used to define functions.

You can use ANNOTATE= data sets to position text labels in a graph, to place symbols or city names on maps, to draw lines between points, and to compose special presentation graphics. You annotate your graphs by using a DATA step to place your commands into an ANNOTATE= data set and then specifying that data set with the ANNOTATE= option in any of the following procedures: GANNO, GCHART, GCONTOUR, GMAP, GPLOT, GSLIDE, and G3D.

For example, if you have an ANNOTATE= data set named GDATA, you can specify

ANNOTATE= data sets now support BY statement processing. The same BY variable(s) must be present in the PROC input data and the ANNOTATE= data set.

ANNOTATE= VARIABLES

FUNCTION

the particular feature you want to activate

X

the X or horizontal coordinate

Y

the Y or vertical coordinate.

ANNOTATE example

FUNCTION Variable

Basic Functions

FUNCTION='BAR'

constructs a rectangle. You can define the color of the fill, the fill pattern, and the edge lines to be drawn.

FUNCTION='COMMENT'

places comments in your program. The text of the comment is ignored when the data set is processed.

FUNCTION='DRAW'

draws a line on the display area. You can define the color, line style, and thickness of the lines to be drawn.

FUNCTION='FRAME'

draws a border around the outside of the defined display area. You can optionally specify a background color for the area of the display enclosed by the frame.

FUNCTION='LABEL'

places text on the display area. You can specify the color, size, font, base angle, and rotation of the characters displayed.

FUNCTION='MOVE'

allows you to move to a specific point on the display area without drawing a line. MOVE is most often used to prepare for a DRAW command or advanced text functions.

FUNCTION='PIE'

draws pie slices on the display area. You can specify the color, fill pattern, arc angle, radius, and edge lines of the slice being drawn.

FUNCTION='POINT'

places a single point at the (X,Y) coordinates using the specified color.

FUNCTION='POLY'

specifies the beginning of a polygon definition. You can define the fill pattern and color, as well as the line type for the outline. The POLY function is used with the POLYCONT function to define and fill areas on the display area.

FUNCTION='POLYCONT'

specifies successive points in the polygon definition in separate continuation observations. The color for the outline is specified in the first POLYCONT observation.

FUNCTION='SYMBOL'

places special symbols on the display area. You can specify the symbol, font, height, and color to be used.

X, Y, Z and Related Variables

X, Y, and Z variables

XC and YC variables

Utility variables

The variables XLAST and YLAST are an internal coordinate pair that track the last values specified for X and Y. Since the (XLAST,YLAST) coordinates are updated internally, you cannot specify values for them. However, these variables can be manipulated by the utility advanced functions documented in the Advanced Functions section. The coordinate pair (XLAST,YLAST) is automatically updated by certain functions and is available for use by other functions that follow.

The coordinate pair (XLSTT,YLSTT) is similar to the (XLAST,YLAST) coordinate pair except that the (XLSTT,YLSTT) pair is only updated by the text-handling function LABEL. Thus, it is possible to maintain two different coordinate pairs: one for text-handling functions and one for nontext handling functions. The (XLSTT,YLSTT) coordinates are updated automatically, and you cannot specify values for them. However, these values can be manipulated indirectly by the advanced functions described in the Advanced Functions section.

X and Y Coordinate Systems

When you specify values for the X and Y variables (and Z variable in PROC G3D), you can also specify the coordinate system to use to place this information on the display area. A coordinate system defines the graphics area, units, and location for displaying Annotate information.

Defining the graphics area

The DATA graphics area is the area that is bordered by the horizontal and/or vertical axis (and in PROC GMAP, the range of map coordinates) specified in a procedure step. For example, if you customize output from PROC GPLOT and specify the DATA coordinate system, the available area is the area enclosed by the axis lines.

The SCREEN graphics area is the entire area available for graphics on the device. When you specify the HSIZE= and VSIZE= options, these values set the physical limit of the SCREEN coordinate system.

The WINDOW graphics area is the same area defined by the SCREEN area minus the amount of space required by TITLE and FOOTNOTE statements.

Defining units for X and Y

Defining location

By combining values to define the area, unit, and location for the Annotate information, you can specify up to twelve coordinate systems. Each unique combination of area, unit, and location definition has been assigned a value that you specify with the XSYS, YSYS, ZSYS, and HSYS variables.

XSYS, YSYS, ZSYS, and HSYS Variables

You can use these coordinate systems in any combination to specify display area locations in the ANNOTATE= data set. The X and Y variables need not be referenced with the same system value. X can be referenced as 'data value' and Y referenced as 'window percentage' in one observation, and both coordinates can be referenced as 'screen value' in the next.

The HSYS variable determines what coordinate system the variable SIZE uses. The coordinate system values specified with XSYS, YSYS, and ZSYS are valid with HSYS.

Additional Variables

ANGLE

is a function-dependent numeric variable. ANGLE can be used with the functions LABEL and PIE. ANGLE is measured in degrees.

COLOR

is a character variable of length 8 that specifies the color used by the function COLOR can be used with the functions BAR, DRAW, FRAME, LABEL, PIE, POLY, and POLYCONT.

GROUP

is a variable used to reference the GCHART GROUP data value. GROUP should be of the same type (character or numeric) as the GCHART variable it references. GROUP is only valid with PROC GCHART. The use of the GROUP variable is similar to that of the X and Y variables when used for data dependent placement.

LINE

is a function-dependent variable that usually specifies the line type. LINE can be used with the functions BAR, DRAW, FRAME, PIE, and POLY.

MIDPNT

is a variable used to reference the GCHART MIDPOINT data value. MIDPNT should be of the same type (character or numeric) as the GCHART variable it references. MIDPNT is only valid with PROC GCHART. The use of the MIDPNT variable is similar to that of the X and Y variables when used for data-dependent placement.

POSITION

is a character variable used to control placement and alignment of a text string. POSITION can be used only with the LABEL function.

ROTATE

is a function-dependent numeric variable. ROTATE can be used with the fuctions LABEL and PIE. ROTATE can be used with the ANGLE variable and is measured in degrees.

SIZE

is a function-dependent numeric variable. SIZE can be used with the functions DRAW, FRAME, LABEL, PIE, and PIEXY. The value of the variable SIZE defaults to 1.00 in all cases.

STYLE

is a function-dependent character variable of length 8 that normally specifies a font or pattern for the function. STYLE can be used with the functions BAR, FRAME, LABEL, PIE, and POLY.

SUBGRP

is a variable used to reference the actual GCHART SUBGROUP data value. SUBGRP should be of the same type (character or numeric) as the GCHART variable it references. SUBGRP is only valid with PROC GCHART. The use of the SUBGRP variable is similar to that of the X and Y variables when used for data-dependent placement.

WHEN

is a sequencing variable that specifies when the function is performed in relation to generating graphics output for the procedure. WHEN can take the values 'A' (AFTER the actual graph is drawn) or 'B' (BEFORE the graph is drawn). A missing value is equivalent to specifying BEFORE. Normally, observations in an ANNOTATE= data set are processed sequentially. If the variable WHEN is used, all those observations with a WHEN value of 'B' are processed first, the graph is then processed (if there is one to be produced), and finally the observations with a WHEN value of 'A' are processed. WHEN should always be a character variable of length 1.

© 1995 Michael Friendly