topic Re: Change line style in ODS graph, Proc Glimmix in SAS Studio

Change line style in ODS graph, Proc Glimmix

jlajla — Tue, 10 Nov 2015 20:40:07 GMT

I'm using SAS Studio & I've plotted a mean response profile for two groups. How do I change the appearance of one line so that the groups can be differentiated in black & white? For instance, how do I make one line dashed, etc.?

Thank you!

ODS graphics on;
Proc Glimmix data=new;
class visit group;
model fact = visit group visit*group;
lsmeans visit*group / plots=(meanplot( join sliceby=trt));
Label fact= 'Mean Score';
Label visit='Follow-Up Visit';
title 'Mean Score at Each Visit';
run;
ODS graphics off; quit;

Re: Change line style in ODS graph, Proc Glimmix

DanH_sas — Tue, 10 Nov 2015 20:54:29 GMT

Your best option might be to just change your ODS style in SAS Studio to use the "journal" style instead of the default (which is probably "HTMLBlue").

Re: Change line style in ODS graph, Proc Glimmix

jlajla — Tue, 10 Nov 2015 21:09:12 GMT

Perfect, thank you!

Do you know if there is another option using coding? I'm still new at SAS so trying to learn...

Re: Change line style in ODS graph, Proc Glimmix

DanH_sas — Tue, 10 Nov 2015 21:20:55 GMT

More than likely, this was your best option. Glimmix (and other stat procs like it) are using GTL templates to define their graphics. It is possible for users to edit these templates to make customizations. However, judging by the Glimmix request, there was probably a grouping variable involved that causing the graph renderer to cycle through the style attributes. For the HTMLBlue style, the line patterns are set to be solid for the first twelve group values, which is why you just saw solid lines. Most other styles (Listing, Statistical, etc.) would change both color and line pattern for each group value. I pointed you to Journal because it is optimized for grayscale output. If you need pure black and white output, you should also give Journal2 a try.

Thanks!

Dan

Re: Change line style in ODS graph, Proc Glimmix

jlajla — Tue, 10 Nov 2015 22:04:29 GMT

great, thank you for the detailed answer!