pointlabel -- needs to be cut short, with int()

[NKormanik]

symbol1 pointlabel = ("#N" h=.5 font=simplex nodropcollisions) value=none;

N is a number of cases.  I'd like to see that number as label in a scatterplot, instead of a dot.

proc gplot data=have;
plot X * Y = 1;  /* Surmised that "= 1" means symbol1 */ 
run;
quit;

Unfortunately N shows in the plot having many decimal places, all 0.

I figured using int() would do the trick, to just show N alone, no superfluous unwanted decimal places.

I tried several ways, but none seem to work.  One approach that does not work is:

"#int(N)"

Another possibility would be to create an ad-hoc variable, N2, that is int(N), and then to use N2 instead of N in the pointlabel.  But I'm not sure how to code such.

Any help toward a solution greatly appreciated.

Nicholas Kormanik

[Rick_SAS]

Would you be willing to use PROC SGPLOT instead?  If so, you can use the MARKERCHAR= option to specify a variable name. Instead of markers, the value of that variable is plotted. You can controls the size and color of the text by using the MARKERCHARATTRS= option:

data Have;
input x y N;
datalines;
1 2 37
3 1 23
5 2 16
4 1  2
5 4  1
3 3 18
;

proc sgplot data=Have;
scatter x=x y=y / markerchar=N markercharattrs=(size=14);
run;

[NKormanik]

Beautiful.  Still, however, all those .000000 after each number in plot.

Got error message when I tried:  

markerchar=int(N)

and

markerchar=(int(N))

Perhaps I should format the N column itself?

Thanks much!

[Jay54]

Yes, apply the format to the column.  If you don't want decimals, use the x.0 format.  If you review the doc, you will see that MARKERCHAR only accepts a variable, not an expression.  If you want to use expressions, use GTL.

[NKormanik]

What's the easiest way to apply that format to the column?

[DanH_sas · Accepted solution]

Use the FORMAT statement:

proc sgplot data=Have;
format N f2.0;
scatter x=x y=y / markerchar=N markercharattrs=(size=14);
run;

[NKormanik]

Excellent.  This approach has the added benefit of not having to affect/rewrite the dataset.

Thanks much to all of you!