data driven coloring in sgplot

[Forecaster]

Hello, I need to create a bullet graph by category. I have actual sales and predicted sales for individual products and the are grouped by category. I would like to use different colors for each category and in the legend show only actual and predicted sales. When I used ssgpanel, the p[roblem I'm facing is that the products in one category is also appearing in other category. Please see below for a reproducible example. I'm using SAS 9.3.

My question is: how can I create a bullet graph by data driven color schemes ?

data sales;

  input cat $ product $ actual predict;

  datalines;

  X A 10 5

  X B 12 15

  X C 15 18

  X D 13 20

  Y E 20 15

  Y F 22 25

  Y G 25 38

  Y H 23 40

  ;

run;

data attrs;

input id $ value $ fillcolor $;

datalines;

my_id X blue

my_id Y red

;

run;

proc sgpanel data = sales dattrmap=attrs;

  PANELBY cat/LAYOUT=  rowlattice;

  hbar product/response = actual group = cat attrid=my_id ;

  hbar product/response = predict group = cat BARWIDTH =0.2 FILLATTRS = (color = black) ;

run;

Products E,F,G and H appear in cat x even though cat x does not have these products vice versa.

[DanH_sas · Accepted solution]

Add UNISCALE=COLUMN to the PANELBY statement, and you should get what you want.

[Forecaster]

Thank you this worked perfectly.