topic How modify Reference line in PROC SGPLOT in Graphics Programming

How modify Reference line in PROC SGPLOT

DmytroYermak — Fri, 04 May 2018 10:21:33 GMT

Hi all,

Could you please help to modify Reference line on the graph below. The issue is the reference line extends beyond timepoints where values of concentration do not exist.

The code that was used is here:

proc sgplot data=abc; 
  where col1 = "&&sub&i";
    styleattrs AXISBREAK=Bracket;
    series x=col3 y=col4/ lineattrs=(pattern=solid color=blue) name="line" markers markerattrs=(symbol=circlefilled color=blue size=6);
	xaxis label='Timepoint (hrs)' type=linear values = (0 to &&xmax&i. by 12)  
    integer valueattrs=(size=8pt) fitpolicy = rotate integer valuesrotate=vertical;
     yaxis label='Semi-log Concentration' 
     type=log logbase=10 logstyle=linear logvtype=expanded values = (0.1 1 10) 
        THRESHOLDMin=1 THRESHOLDMAX=1;
	refline col7  /axis=y lineattrs=(color=red) name = "line1" legendlabel = "Mean X-response";
    keylegend "line" "line1" / location=inside position=topright across=1 noborder;

run;

Do any options for refline exist to modify it?

Thank you!

Re: How modify Reference line in PROC SGPLOT

RW9 — Fri, 04 May 2018 10:47:22 GMT

As far as I am aware the refline will fill across the width or height of the axis - it is a reference line at the end of the day. What you could do is just overlay a series plot of x=0, y=col7 and x=60, y=col7.

Re: How modify Reference line in PROC SGPLOT

DmytroYermak — Fri, 04 May 2018 11:00:02 GMT

The matter is that the timepoint "72 hours" exists with <null> value of concentration.

Re: How modify Reference line in PROC SGPLOT

RW9 — Fri, 04 May 2018 11:08:58 GMT

Just add it to your where clause then:

  where col1 = "&&sub&i" and col4 ne .;

Also note, you don't need the macro looping - which I can infer from the &&sub&i, you can use by grouping:

Re: How modify Reference line in PROC SGPLOT

ballardw — Fri, 04 May 2018 17:30:56 GMT

@DmytroYermak wrote:
The matter is that the timepoint "72 hours" exists with <null> value of concentration.

?????

I always get confused when people reference "null" in SAS discussions since SAS uses the term "missing".

And why no complaint about the reference line showing at x=0?

Old school solution would be create another plot series with matching x values that start and end with corresponding values of y with the refline value.

Of course if the refline is changing for each level of your macro variable that could lead to some mildly entertaining code.