SGPLOT - Attrmap and Proc Format not working together

[DME790]

Hi All,

I'm trying to use Attrmap and a proc format but they seem to cancel each other over.

If I run the code without the format Simplification typeFmt. I get the green from the Attrmap - as expected -  and legend = 0 (pic 1).

If I run with format Simplification typeFmt. I loose the formatting from Attrmap and the legend   = Pre trial (pic2) 

Is there a way I can have both  - Attrmap and the proc format work together. Green columns from the Attrmap and a legend shows Pre trial and Post Trial. 

Data Attrmap;
	Retain id "myid";
	input value $ fillcolor $8.;
	datalines;
	0 VIBG
	1 BIBG
	;
run;

proc format;
	value typeFmt
		0 = "Pre Trial"
		1 = "Post Trial" 
	;
	ods graphics on / width=9in height=5in;

Proc SGPLOT data=work.DSCMax noborder dattrmap=attrmap;
	*/
	format Simplification typeFmt.;
	VBAR EVENT_DT / Response=Max_answer_tm Group=Simplification attrid=myid nooutline seglabel seglabelattrs=(weight=Bold color=WHITE size=8 );
	yaxis display=(noline noticks) grid values=(0 to 3600 by 300) offsetmin=0 label='mm:ss';
RUN;

ODS graphics OFF;

Apologies if the test data doesn't run - this is what it looks like though

data test;
	Input CD EVENT_DT Group Simplification MAX_ANSWER_TM;
	Datalines;
TSDa 	2018-07-06	DSC	0	38:21
TSDa	2018-07-09	DSC	0	9:41
TSDa	2018-07-10	DSC	0	23:31
TSDa	2018-07-11	DSC	0	14:47
TSDa	2018-07-12	DSC	0	17:40
TSDa	2018-07-13	DSC	0	20:34
TSDa	2018-07-16	DSC	0	8:55
TSDa	2018-07-17	DSC	0	22:17
TSDa	2018-07-18	DSC	0	16:42
TSDa	2018-07-19	DSC	0	10:31
TSDa	2018-07-20	DSC	0	11:12
TSDa	2018-07-23	DSC	0	10:13
TSDa	2018-07-24	DSC	0	6:37
TSDa	2018-07-25	DSC	0	17:00
TSDa	2018-07-26	DSC	0	7:11
TSDa	2018-07-27	DSC	0	10:43
;
RUN;

Pic 1                                                                                                 Pic 2

[FreelanceReinh · Accepted solution]

Hi @DME790,

According to the documentation: "If the group value is formatted, then the text-string in the attribute map data set must contain the formatted value."

So, just modify variable VALUE in dataset ATTRMAP:

data test;
input cd $ event_dt :yymmdd. group $ Simplification max_answer_tm :time.;
max_answer_tm=max_answer_tm/60;
format event_dt yymmdd10. max_answer_tm mmss.;
label event_dt='Event Date';
datalines;
TSDa 2018-07-06 DSC 0 38:21
TSDa 2018-07-09 DSC 0 9:41
TSDa 2018-07-10 DSC 0 23:31
TSDa 2018-07-11 DSC 0 14:47
TSDa 2018-07-12 DSC 0 17:40
TSDa 2018-07-13 DSC 0 20:34
TSDa 2018-07-16 DSC 0 8:55
TSDa 2018-07-17 DSC 0 22:17
TSDa 2018-07-18 DSC 0 16:42
TSDa 2018-07-19 DSC 0 10:31
TSDa 2018-07-20 DSC 0 11:12
TSDa 2018-07-23 DSC 0 10:13
TSDa 2018-07-24 DSC 0 6:37
TSDa 2018-07-25 DSC 0 17:00
TSDa 2018-07-26 DSC 0 7:11
TSDa 2018-07-27 DSC 0 10:43
TSDa 2018-07-27 DSC 1 -0:01
; /* last obs. added to force "Post Trial" into the legend */

data attrmap;
retain id "myid";
input value :&$10. fillcolor $8.;
datalines;
Pre Trial   VIBG
Post Trial  BIBG
;

ods graphics on / width=9in height=5in;

proc sgplot data=test noborder dattrmap=attrmap;
format Simplification typeFmt.;
vbar event_dt / response=max_answer_tm group=Simplification attrid=myid nooutline seglabel seglabelattrs=(weight=bold color=white size=8 );
yaxis display=(noline noticks) grid values=(0 to 3600 by 300) offsetmin=0 label='mm:ss';
run;

ods graphics off;

If you don't need "Post Trial" in the legend if only "Pre Trial" values occur in the graph, just delete the dummy observation that I've added.

[DME790]

Thanks @FreelanceReinh

Works a treat

Cheers

Dean