08-04-2015 04:02 AM
I have a dataset with 4 variables
the value of z is spaced in the interval from 0 to 1;
I made a dataset grided by g3grid, which works perfectly.
I want now to plot for each id (ptid) a 3D plot with the same z contour levels (logs). However in some of my graps, proc gcontour overrules my clevels. It is because for the specific id (ptid) the data may not contain values in some of the range. But I dont care. Then proc gcontour should simply plot with the clevels I have chosen, if no data fit then it should plot no contours at all. How can I stop proc contour trying to be intelligent and let me decide the clevels.
proc gcontour data=grided; by ptid;
levels= 0.2 0.4 0.6 0.8
clevels=white red orange yellow green blue black
2 weeks ago
I'm not too familiar with Proc Gcontour, and what its limitations might be ... but if you're using device=activex (or device=java), try it with device=png and see if it has different functionality.
2 weeks ago
Try this and see what you get:
proc template; define statgraph contourplotparm; begingraph; entrytitle "My Contour"; layout overlay / xaxisopts=(offsetmin=0 offsetmax=0 linearopts=(thresholdmin=0 thresholdmax=0)) yaxisopts=(offsetmin=0 offsetmax=0 linearopts=(thresholdmin=0 thresholdmax=0)); contourplotparm x=sequence y=loghz z=logs / contourtype=fill levels=(0.2 0.4 0.6 0.8) name='Contour'
colormodel=(white red orange yellow green blue black); continuouslegend 'Contour' / title="My Levels"; endlayout; endgraph; end; run; proc sgrender data=grided template=contourplotparm; by ptid; run;