SGPANEL and COLAXIS

[MeredithG]

It seems like my COLAXIS statement is being ignored:

proc sgpanel data=long noautolegend;

     panelby group time/onepanel novarname sparse columns=2 rows=3 ;

     vbar dose;

     colaxis values=(0 to 120 by 30);

run;

I get an x-axis with a bunch of unreadable cluttered dose numbers.

[PGStats]

I think that vbar implies a discrete X axis. Try FITPOLICY=THIN in your rowaxis statement.

PG

[MeredithG]

I have, it doesn't make them thin enough to be readable. I may just make a new ordinal variable and use that instead.

Edit: I'm going to use a DENSITY plot. I have a bunch of other panels with discrete axes and I wanted them to look uniform, but it's not right for this variable.
Thank you!

[PGStats]

If dose is a continuous variable, taking all sorts of values, then perhaps histograms would be preferable to vbar graphs. - PG

[DanH_sas]

or NEEDLE plots.