I am doing a multicell plot in GTL with some barcharts and what to control the colors with an attributemap.  The grouping within each of the cell plots are different . To illustrate the "problem" I have played with sashelp.cars where the two plots use car type for group, the other drive train.    In the first cell we have a barchart of engine size by car type.  In the second cell we have a barchart of engine size by drive train.  The plot itself is okay as such but I want to control the colors with an attributemap. I also want to show a legend group item if data is available for that group. Have tried several work around by can not get it to work......    Here is the code: data attrmap; retain ID "cars" linecolor "black"; input VALUE $ 1-9 FILLCOLOR $ 10-19 FILLPATTERN $ 20-29 ; *--------1---------2---------3---------4---------5---------6---------7---------8---------9; datalines; Hybrid black solid SUV yellow solid Sedan blue solid Sports vivid solid Truck hue solid Wagon orange solid All red l1 Front green l2 Rear yellow l3 ; run; proc template; define statgraph multicell; begingraph; layout lattice / rowgutter=10px columngutter=10px; layout overlay; entry "Car type" / valign=top; barchart x=origin y=engine/ group=type name="legend_1" ; discretelegend "legend_1" /displayclipped=true autoitemsize=true LOCATION=outside; endlayout; layout overlay; entry "Car drivetrain" / valign=top; barchart x=origin y=engine/ group=drivetrain name="legend_2" ; discretelegend "legend_2" /displayclipped=true autoitemsize=true LOCATION=outside; endlayout; endlayout; endgraph; end; run; proc summary data=sashelp.cars; class origin type drivetrain; var enginesize; output mean=engine out=data4plot; run; proc sgrender data=data4plot template=multicell dattrmap=work.attrmap; run;   ... View more

Trying to get a a the symbols (scatter) and line (series) legends be merge together but can not get the group label on. Any ideas on how to solve this in a dynamic way?  So in the below simple example I want to have F and M included in the legend label.   proc sort data=sashelp.class out=class; by age; proc sgplot data=class; scatter x=age y=height / group=sex; series x=age y=height / group=sex; legenditem type=markerline name="symbol"/ lineattrs=GraphData1 markerattrs=GraphData1; legenditem type=markerline name="line" / lineattrs=GraphData2 markerattrs=GraphData2; keylegend "symbol" "line" /title="Sex"; run;     ... View more

Hi folks I want some observations (obs)  by group (place) and time (timepoint)  data have; input place $ timepoint obs; datalines; a 1.1 0 a 1.2 0 a 1.3 1 a 1.4 0 a 1.5 0 b 1.1 0 b 1.2 1 b 1.3 0 b 1.4 2 b 1.5 0 ; run; data want; set have; ... run; What I want is to retain the last obs by place to the dataset looks like this Any idea on how to do this ?  ... View more

I am doing an template for an AE dotplot for clinical trials. The text string for a preferred term is shown on the y axis. Often the text string is very long why I want have a line break so the term will be on two lines (the y axis tick mark label). I have tried to get force a line shift in the text string by setting in the unicode for a line break  - (*ESC*){unicode '000a'x} - but that does not work.  I am doing the plot in GTL   proc template; define statgraph aedotplot; begingraph ; layout lattice /columns=5 columnweights = (0.4 0.1 0.1 0.2 0.2) rowdatarange=unionall columngutter=0 ; layout overlay / walldisplay=none border=false x2axisopts=(display=(tickvalues ) displaysecondary=(line)) yaxisopts=(linearopts=(tickvaluesequence=(start=1 end=&count increment=1)) display=none displaysecondary=none ); /*--Draw alternating bands using referenceline--*/ referenceline y = display2nd/ lineattrs=(thickness=&RefThickness.PCT) datatransparency=0.9; /* left colum with text box */ scatterplot y=display x=display_label / markercharacter=display MARKERCHARACTERPOSITION=right yaxis=Y xaxis=X2 markerattrs=(size=0) includemissinggroup=true; endlayout; ......     Any ideas ?            ... View more

Trying to get the geometric mean with  SE, STD and CLM on some clinical data and want the outputs stats to be saved as a SAS datafile for further processing.    Was thinking of using proc surveymeans as I want the variance of the geometric mean to be estimated using the Taylor series.  How do I get the relevant stats in the output dataset ?   Below code only gives me the SE, STD and CLM to the arithmetic mean - NOT the geometric mean.    Somebody that can help a newbie  to surveymeans ?      A small example:  proc sort data=sashelp.class out=test;  by sex; run;   proc surveymeans data=test alpha=0.05 geomean std stderr clm ;  by sex;  var age;  ods output    Statistics = stats_ds; run;   ... View more