One of most powerful
features of the GTL is the syntax built around hierarchical statement
blocks called “layouts.” The outermost layout block
determines
-
The overall organization of the
graph—whether it uses a single-cell or a multi-cell display.
-
What statements are allowed in
the block. Generally, layout blocks can contain plots, lines of text,
a legend, or even another layout.
-
How the contained statements interact.
Outermost Layouts in GTL
|
|
|
General purpose layout
for displaying 2-D plots in a single-cell.
|
|
Layout for displaying
3-D plots in a single-cell.
|
|
Specialized OVERLAY
with equated axes.
|
|
General purpose layout
for displaying single-cell graphs that does not use axes.
|
|
Basic grid of plots.
All cells are independent.
|
|
Advanced multi-cell
layout. Axes can be shared across columns or rows and be external
to grid. Many grid labeling and alignment features.
|
|
Generates a classification
panel from the values of 1 or 2 classifiers.
|
|
Generates a classification
panel from the values of n classifiers.
|
|
Specialized layout for
creating a compound legend that contains multiple discrete legends.
|
For example, the following
graph is a two-cell graph produced using the LAYOUT LATTICE statement
as the outermost template in the layout.
The LAYOUT LATTICE statement
is typically used to create a multi-cell layout of plots that are
aligned across columns and rows. In the following template, which
produced the graph, plot statements are specified within nested LAYOUT
OVERLAY statements. Thus, the LATTICE automatically aligns the plot
areas and tick display areas in the plots. The LATTICE layout is a
good layout to choose when you want to compare the results of related
plots.
proc template;
define statgraph lattice;
begingraph;
entrytitle "Car Performance Profile";
layout lattice / border=true pad=10 opaque=true
rows=1 columns=2 columngutter=3;
layout overlay;
scatterplot x=horsepower y=mpg_city /
group=origin name="cars";
regressionPlot x=horsepower y=mpg_city / degree=2;
endlayout;
layout overlay;
scatterplot x=weight y=mpg_city / group=origin;
regressionPlot x=weight y=mpg_city / degree=2;
endlayout;
sidebar;
discretelegend "cars";
endsidebar;
endlayout;
endgraph;
end;
run;
For detailed information
about each layout, see the chapter for that layout type.