Glossary

anti-aliasing

a rendering technique for improving the appearance of text and curved lines in a graph by blurring the jagged edges normally present. The degree of improvement is relative to the nature of the graphical content (for example, vertical and horizontal lines do not benefit from anti-aliasing). Extra processing is required to perform anti-aliasing.

attribute bundle

a common collection of visual properties associated with a graphical primitive such as a line, marker, or text. For example, all lines have visual properties of pattern, thickness, and color. All markers have visual properties of symbol, size, weight, and color. Attribute bundles can be associated with style elements in order to indirectly assign visual properties.

axis

a graphical element used to locate or identify the values of other graphical elements, such as points or bars. An axis consists of an axis line with tick marks, tick values, and a descriptive label. Not all parts of the axis need to be displayed. An axis is typically an interval axis (linear, log, or time) or a discrete axis. A two-dimensional graph can have up to four independent axes: X, Y, X2, and Y2. A three-dimensional graph has three independent axes: X, Y, and Z.

axis offset

the gaps that normally appear at the ends of an axis line. The gaps enable markers, bars, and other graphic primitives that are drawn at extreme data values to be rendered without clipping. An offset can also be used to add extra space between an axis line and visual elements in the graph. An offset distance is expressed as a value from 0 to 1, which represents a percentage of axis length to the axis data minimum or axis data maximum. An offset can be specified for either end of any axis.

axis threshold

a numerical bias from 0 to 1 that determines whether an extra tick is added at either end of a non-discrete, interval axis. If the minimum and maximum thresholds are set to 0, then no ticks are added beyond the actual data range. If both minimum and maximum thresholds are set to 1, then the data range is completely bounded by the first and last ticks.

axis tick

a short line segment perpendicular to the axis line. A tick can cross the axis line, or be drawn from the axis inside or outside the wall.

axis tick value

a formatted data value represented by a tick.

axis type

a keyword denoting axis functionality. For example, the axis type of interval axes can be LINEAR, TIME, or LOG. The axis type of a discrete axis is DISCRETE.

axis viewport

the range of values displayed on an interval axis. This range can be larger or smaller than the actual data range of the axis. An axis viewport that is larger than the data range effectively zooms out from the plot or plots. An axis viewport that is smaller than the data range zooms in on the plot or plots.

band plot

a plot that draws a horizontal band using two Y values for each X value, or that draws a vertical band using two X values for each Y value. A band plot is typically used to show confidence, error, prediction, or control limits. The points on the upper and lower band boundaries can be joined to create two outlines, or the area between the boundaries can be filled.

binned data

data that has been summarized or transformed in some way to facilitate its rendering by a parameterized plot. Continuous numeric data is typically binned by setting a bin width (interval size) and then computing the number of bins, or by setting the number of bins and computing the bin width. A histogram is often used to represent binned data.

bins

numeric intervals into which continuous numeric data can be categorized.

block plot

a plot that displays one or more rectangles (blocks) along an axis, where each rectangle identifies a block of consecutive observations having the same value for a specified block variable. The first block begins at the start of the axis (mapped to the values of a specified variable), and represents the first observation s block value, and continues through consecutive observations having the same block value. A change in the block variable s value ends the first block and starts the second, which continues through consecutive observations until the block value changes again. The last block extends to the end of the axis.

border

the outermost outline of a graph, a layout container, or a graphical element (for example, a legend or text).

category variable

a classification variable with a finite number of distinct (discrete) values. These variables are typically used to split data into subsets. For example, in a bar chart, each unique value is displayed as a bar on a DISCRETE axis.

cell

in ODS graphics, a distinct rectangular subregion of a graph that can contain plots, text, or legends. For example, a graph defined with the OVERLAY layout contains only one cell, whereas a graph defined with the GRIDDED layout can contain several cells.

cell block

a block beginning with a CELL statement and ending with an ENDCELL statement that defines the graphical content of a cell. The cell block is available only within a LATTICE layout.

cell header

a graphical element (typically text or a legend) that is aligned at the top of a cell and provides information about the cell contents. A cell header is defined within a cell block, which is available only within a LATTICE layout.

child block

a block that is contained within another block when two or more blocks are nested. For example, a CELLHEADER block is always a child of a CELL block.

classification level

for a single classification variable, each unique value is regarded as a classification level. For two or more variables, a classification level is one of the unique combinations (crossings) of the unique values of each variable. For example, if three variables have four, two, and three distinct values, there are 24 classification levels.

classification panel

a multi-cell graph created by a layout, such as DATAPANEL, in which the number of cells is determined by the number of classification levels of one or more classification variables. Each cell displays a common plot based on subsets of the input data.

classification variable

a variable whose values are used to group (or classify) the observations in a data set into different groups that are meaningful for analysis. A classification variable can have either character or numeric values. Classification variables include group, subgroup, category, and BY variables.

clip

to truncate a plot or graphical element (such as a line, marker, or band) when it reaches a boundary such as a plot wall.

column

a vertical component of a table. Each column has a unique name, contains data of a specific type, and has certain attributes. A column is analogous to a variable in SAS terminology. In the Graph Template Language a column can also be a set of layout cells, stacked vertically and sharing the same alignment.

column axis

an external axis appearing above or below a column of cells and serving as a common reference for the column of a multi-cell layout, such as a LATTICE, DATAPANEL, or DATALATTICE layout.

column gutter

the space between columns of cells in a multi-cell layout.

column header

text that labels the column contents in a multi-cell layout. This text can be aligned above or below the cells in a column. In a LATTICE layout, the column header is not restricted to text (it can contain a plot or a legend, for example).

column major order

an order for populating cells of a layout or entries in a legend when the number of rows is specified. By default, cells or entries are filled starting from the top left and moving down. When the bottom row of the first column is filled, a new column begins filling to the right of the previous column, and so on until all content items have been placed in cells or entries. There might be empty cells or entries in the last column.

column weight

in a LATTICE layout, the proportion of width allotted to a specific column of the layout. The sum of all column weights is 1.

computed plot

a plot in which input data is internally summarized or otherwise transformed to create new data that is actually rendered by the plot. Examples of computed plot statements are BARCHART, BOXPLOT, HISTOGRAM, ELLIPSE, and REGRESSIONPLOT.

conditional logic

syntax that enables one set of statements or an optional alternate set of statements to execute at run time. In the Graph Template Language, an IF/ENDIF block defines conditional logic: IF (condition) statements; ELSE statements; ENDIF; The ELSE statement is not required.

continuous legend

a legend that shows a mapping between a color ramp or color segments and corresponding numeric values. Plots that support a COLORMODEL= option can use this type of legend.

crossing

a combination of the unique values of one or more classification variables. See also classification level.

cube

in three-dimensional graphics, the outlines formed by the intersection of three pairs of parallel planes; each pair is orthogonal to the primary X, Y, and Z axes. The display of the cube is optional.

data object

a transient version of a SAS data set created by ODS. When an input SAS data set is bound to a compiled graph template, an ODS data object is created, based on all the columns requested in the template definition and any new columns that have been directly or indirectly computed. A data object can persist when used with the ODS OUTPUT statement.

data tip

data or other detailed information that is displayed when a user positions a mouse pointer over an element in a graph. For example, a data tip typically displays the data value that is represented by a bar, a plot point, or some other element.

define block

in the TEMPLATE procedure, a define block (beginning with a DEFINE statement and ending with an END statement) creates various types of templates, including STATGRAPH, STYLE, and TABLE.

dependent plot

a plot that cannot be rendered by itself. Dependent plots must be overlaid with a stand-alone plot. Dependent plots do not provide data ranges to establish axes. REFERENCELINE, DROPLINE, and LINEPARM statements produce dependent plots. See also stand-alone plot.

dependent variable

a variable that is observed to change in response to the independent variables. In a function y=f(x), the value of the dependent variable y is a function of the independent variable x.

design size

the intended size of a graph that is specified in the graph template definition. The DESIGNHEIGHT and DESIGNWIDTH options of the BEGINGRAPH statement set the intended height and width, which are used to determine the scale factors when the graph is resized. The intended height and width are used unless overridden by the ODS Graphics statement HEIGHT or WIDTH options when the template is executed.

device-based graphic

a graph created with traditional SAS/GRAPH software that requires DEVICE specification. ODS graphics (template-based graphics) do not use device technology.

discrete axis

an axis for categorical data values. The distance between ticks has no significance. A bar chart always has a discrete axis.

discrete legend

a legend that provides values or descriptive information about graphical elements in a grouped or overlaid plot.

dots per inch

a measure of the graph resolution by its dot density. Short form: DPI.

DPI

See dots per inch.

drop line

a line drawn from a point in the plot area perpendicular to an axis.

dynamic variable

a variable defined in a template with the DYNAMIC statement that can be initialized at template run time.

equated axes

in two-dimensional plots, axes that use the same drawing scale (ratio of display distance to data interval) on both axes. For example, an interval of 2 on the X axis maps to the same display distance as an interval of 2 on the Y axis. The aspect ratio of the plot display equals the aspect ratio of the plot data. In other words, a 45-degree slope in data will be represented by a 45-degree slope in the display. Equated axes are always of TYPE=LINEAR. The number of intervals displayed on each axis does not have to be the same.

external axis

an axis that is outside all cells of a layout. An external axis represents a common scale for all plots in a row or column of a multi-cell layout.

fill

to apply a color within a bounded area. Many plots, such bar charts and band plots, have bounded areas that can be filled or unfilled. When filled, a color is applied. When unfilled, the areas are transparent.

fit policy

one of several algorithms for avoiding tick-value collision when space allotted to a predefined area does not permit all the text to fit. For example, an axis might have a THIN policy that eliminates the display of tick values for alternate ticks. A ROTATE policy would turn the tick values at a 45-degree angle. A TRUNCATE policy would truncate all long tick values to a fixed length and add an ellipsis (. . .) at the end to imply truncation. A STAGGER policy would create two rows of tick values with consecutive tick values alternating between rows. A compound policy such as STAGGERROTATE could be used to automatically choose the best fit policy for the situation.

footnote area

the region below the graph area where text produced by ENTRYFOOTNOTE statements appears.

frequency variable

in an input data set, a non-negative and non-zero integer variable that represents the frequency of occurrence of the current observation, essentially treating the data set as if each observation appeared <userSuppliedValue>n</userSuppliedValue> times, where <userSuppliedValue>n</userSuppliedValue> is the value of the FREQ variable for the observation.

fringe plot

a plot consisting of short, equal-length line segments drawn from and perpendicular to an axis. Each observation of a numeric variable corresponds to the location for a line segment.

glyph

a letter, character, punctuation mark, pictogram, or symbol that is rendered in the context of some written language. A typeface (font) consists of a coordinated set of glyphs. See also Unicode.

graph

a visualization created by SAS/GRAPH software. A graph that is created by the ODS Graphics system can contain titles, footnotes, legends, and one or more cells, and is typically saved as an image or an SGE file. A generic term for final graphical output without regard to content or format.

graph(ical) area

the region where the visualization displays between the title area and footnote area. The graphical area consists of one or more cells. See also title area and footnote area.

graph(ics) template

See ODS Graphics template.

grid

rows and columns of a multi-cell layout.

gridded data

input that contains at least three numeric variables. Two of the variables are treated as X and Y variables and the third variable Z is treated as if it were a function of X and Y. The X and Y variable values occur at uniformly spaced intervals (although the size and number of intervals might be different for X and Y). All X,Y pairs are unique, and Z values are interpolated so that every X,Y pair has a Z value. Raw data that has at least three numeric variables can be converted to gridded data with the G3GRID procedure (in SAS/GRAPH). The procedure offers both bivariate and spline interpolation methods for computing Z values.

group index

a numeric variable with positive integer values that correspond to values of a group variable. The index values are used to associate GraphData1 GraphData<userSuppliedValue>N</userSuppliedValue> style elements with group values.

group variable

a variable in the input data set used to categorize chart variable values into groups. A group variable enables the data for each distinct group value to be rendered in a visually different manner. For example, a grouped scatter plot displays a distinct marker and color for each group value.

gutter

the space between columns or rows of cells in a multi-cell layout.

image format

a file format that displays a graphical representation. PNG, GIF, TIFF, and JPEG are examples of image formats, each with different characteristics.

image map

in an HTML file, the information contained in the map tag. This information can be used by a browser to display data tips on the image.

independent variable

a variable that persists and affects dependent variables. For a function y=f(x), the value of the dependent variable y is a function of the independent variable x. For example in a REGRESSIONPLOT statement, the X= variable is the independent variable.

inset

a graphical element such as a legend, line of text, or a table of text that is embedded inside of a graph's plot area.

interval axis

an axis where the distance between tick marks represents monotonically increasing or decreasing numeric units of some scale (like a ruler). The standard interval axis is called a LINEAR axis. Specialized interval axes include a TIME axis and a LOG axis.

layout

a generic term for a rectangular container that lays out the positions and sizes of its child components.

layout block

a block beginning with a LAYOUT statement and ending with an ENDLAYOUT statement.

layout grid

a multi-cell layout arranged as a grid of cells in rows and columns.

layout type

a keyword indicating the functionality of the layout. For example OVERLAY, LATTICE, and DATAPANEL are layout types.

legend entry

a combination of a graphical element such as a marker or line along with text describing the value or use of the graphical element. A discrete legend can have several legend entries.

legend title

text that explains how to interpret the legend.

line property

a value that defines the pattern, thickness, or color of a line. By default, the value for a line property is derived from a style element in the current style.

linear axis

an interval axis with ticks placed on a linear scale.

loess plot

a curved line showing a loess fit for a set of points.

log axis

an interval axis with ticks placed on a log scale with a base of 10, 2, or e. A log axis can show ticks 1) that are uniformly spaced representing integer powers of the base, 2) that are uniformly spaced representing the expanded integer powers of the base, or 3) that are uniformly spaced in the data domain (100, 200, 300), but may be shown in a non-uniform spacing on the axis.

macro variable reference

in a template definition, a reference to a macro variable that has been declared with MVAR or NMVAR statements. These references are meant to be resolved at template run time and should not be preceded with an ampersand. If a standard macro variable reference (a name preceded with an ampersand) appears in a template definition, it is resolved at template compile time.

marker

a symbol such as a dot, a cross, a diamond, a circle, or a triangle that is used to indicate the location of, or annotate, a data point in a plot or graph.

marker property

a value that defines the symbol used as a marker, or its size, weight, or color. By default, the value for a marker property is derived from a style element in the current style.

multi-cell layout

a layout that supports a rectangular grid of cells, each of which can contain a graphical element, such as a plot, a legend, a nested layout, and so on.

nested layout

a layout block that appears within the scope of another layout block.

ODS

See Output Delivery System.

ODS Graphics

an extension to ODS that is used to create analytical graphs using the Graph Template Language.

ODS Graphics Editor

an interactive application that can be used to edit and annotate ODS Graphics output.

ODS Template

a description of how output should appear when it is formatted. ODS templates are stored as compiled entries in a template store (item store). Common template types include STATGRAPH, STYLE, CROSSTABS, TAGSET, and TABLE.

opaque

a property of a background. Opaque backgrounds are filled with a color. Non-opaque backgrounds are transparent.

outlier

a data point that differs from the general trend of the data by more than is expected by chance alone. An outlier might be an erroneous data point or one that is not from the same sampling model as the rest of the data.

Output Delivery System

a component of SAS software that can produce output in a variety of formats such as markup languages (HTML, XML), PDF, listing, RTF, PostScript, and SAS data sets. Short form: ODS.

overlay

a plot that can be superimposed on another plot when specified within an overlay-type layout. A common overlay combination is a fit line on a scatter plot.

overlay layout

a type of layout that supports the superimposition of graphical components, such as plots, legends, and nested layouts.

padding

space added inside the border of a graphical component, such as a layout or a legend.

panel

a graph with multiple cells.

parameterized plot

a non-computed plot that requires parameterized data. The Graph Template Language offers several plots in both computed and parameterized versions, for example, BARCHART and BARCHARTPARM. Some computed plots such as REGRESSIONPLOT can be emulated with a SERIESPLOT if the input data represented points on a fit line.

parent block

when two or more blocks are nested, any layout block that contains one or more layout blocks is a parent of the contained blocks.

plot

a visual representation of data such as a scatter plot, a series line, or a histogram. In the ODS Graphics context, plot is a generic term for the graphical element or elements drawn by a plot statement. Multiple plots can be overlaid in a cell to create a graph.

plot area

the space, bounded by the axes, where a visual representation of data, such as a scatter plot, a series line, or a histogram, is drawn.

plot type

a plot family such as bar chart (which would include horizontal, vertical, and grouped bar charts), or a classification scheme for plots based on some useful criteria, such as whether the plots are computed or parameterized.

primary axis

the X or Y axis contrasted to the X2 or Y2 secondary axis.

primary plot

the plot in an overlay that determines axis features, such as axis type and axis label.

prototype layout

an overlay plot composite that appears in each cell of a classification panel. Each instance of the prototype represents a different subset (classification level) of the data.

regression plot

a straight or curved line showing a linear or higher order regression fit for a set of points.

required argument

a variable or constant that must be specified in order to evaluate an expression or render a plot, legend, text, or a layout. For example, a scatter plot has two required arguments: X=column and Y=column.

response variable

See dependent variable.

rich text

a generic term for text that can have different font characteristics (color, family, size, weight, style) on a character-by-character basis and can also be used as a superscript or subscript. All text statements in GTL support rich text.

role

a description of the purpose that a variable serves in a plot. For example, a series plot has predefined roles named for X , Y, GROUP, and CURVELABEL.

row

a set of layout cells that are side-by-side and share the same alignment.

row axis

an external axis appearing on the left or right of a row of cells in a multi-cell layout.

row gutter

space between rows of cells of a multi-cell layout.

row header

typically, the text that identifies the row contents in a multi-cell layout. This text can be aligned to the right or left of the cells in a row. The row header is not restricted to text (it can contain a plot or a legend, for example).

row major order

an order for populating cells of a layout or entries of a legend when the number of columns is specified. For example, in the default case: Start at the top left and fill cells or entries left-to-right. When the right-most column is filled, begin a new row below the previous row. Continue this until all content items have been placed in cells or entries. There might be empty cells/entries in the last row.

row weight

in a LATTICE layout, the proportion of height allotted to a specific row of the layout. The sum of all row weights is 1.

secondary axis

the X2 or Y2 axis as contrasted to the X or Y primary axis.

SGE file

a file created in the ODS Graphics environment that contains an editable graph. Such files have a .SGE file extension and can be edited only with the ODS Graphics Editor. You can edit SGE files from the SAS Results window or by opening the SGE file from within the ODS Graphics editor.

sidebar

an area of certain multi-cell layouts external to the grid of cells where text or other graphical elements can appear. The LATTICE, DATAPANEL, and DATALATTICE layout support four sidebar areas (TOP, BOTTOM, LEFT, and RIGHT).

single-cell layout

a layout type that supports only one cell. The OVERLAY, OVERLAY3D, and OVERLAYEQUATED layouts are examples of single-cell layouts.

sparse data

in classification panels with two or more classifiers, some crossings of the classification values might not be present in the input data. Such input data is called sparse data. By default, a DATAPANEL layout does not generate cells for sparse data, but if requested, it can produce empty cells as place holders for the non-existent crossings.

stand-alone plot

a plot that has its own data range and can therefore appear by itself in a layout.

statement block

a group of statements that has both a logical beginning and ending statement. For example, a LAYOUT statement along with its ENDLAYOUT statement and all contained statements are a block. Some blocks can be nested within other blocks.

style

an ODS template that can be used to control the visual aspects (colors, fonts, lines, markers, and so on) of a graph. A style consists of many style elements, and each style element consists of style attributes.

style attribute

a visual property such as a color, line pattern, or font property that has a reserved name. For example, COLOR, FONTFAMILY, FONTSIZE, FONTWEIGHT, and FONTSTYLE are all attributes of style elements such as GraphTitleText, GraphLabelText, and so on. Style attributes are collectively referenced by a style element within a style definition.

style element

a named collection of style attributes that affects specific parts of ODS output. For example, the GraphTitleText style element specifies the color and font properties of title text and possibly other text in the graph.

style reference

a part of the Graph Template Language syntax that indicates the current value of a specific attribute of a specific style element. For example, SIZE=GraphTitleText:FontSize means to assign to SIZE the value of the FontSize attribute of the GraphTitleText style element from the current style.

surface plot

a three-dimensional graph that displays values of a vertical Z variable based on gridded X and Y variables.

template compile time

the phase when the source program of a template definition is submitted. The syntax of the definition is evaluated for correctness. If no errors are detected, the definition is converted to a binary format and stored for later access.

template definition

the TEMPLATE procedure source program that creates a template. A template definition can be generated from a compiled template. Also called the template source.

template run time

the actions performed when a compiled template is bound to a data object and then rendered to produce a graph. Run-time errors can occur that prevent a graph from being produced.

template source

See template definition.

template store

an item store that contains definitions that were created by the TEMPLATE procedure. Definitions that SAS provides are in the item store Sashelp.Tmplmst. You can store definitions that you create in any template store to which you have write access.

template-based graphic

SAS/GRAPH output where a compiled ODS template of the type STATGRAPH is used to produce graphical output, that is, a graph produced within the ODS graphics environment as opposed to a graph produced in the traditional device-based environment.

text properties

a common set of characteristics that can be specified for any text string: COLOR, FAMILY, SIZE, WEIGHT, and STYLE. By default, values for these properties are derived from a style element in the current style.

time axis

an axis type that displays only SAS date, time, or datetime values. Axis tick value increments can be specified as time or date intervals, such as MINUTE, HOUR, DAY, WEEK, MONTH, QUARTER, or YEAR.

title area

the region above the graph area where text produced by ENTRYTITLE statements appears.

transparency

the degree to which a graphic element (such as a marker or filled area) is opaque or transparent. Transparency is indicated with a number from 0 (completely opaque) to 1 (completely transparent).

Unicode

an encoding system that provides a single comprehensive mapping of all characters (glyphs) in all languages to unique numeric values called code points.

viewport

See axis viewport.

wall

the area bounded by orthogonal axis pairs. In two-dimensional graphs, there is one wall bounded by the XY axes. In three-dimensional graphs, there are three walls, bounded by the XY, YZ, and XZ axes. A wall has an optional outline and can be opaque or transparent.

weight variable

a positive numeric variable in the input data set that represents a weight to be applied to the current observation.