how to control label for the x-axis

[bijayadhikar]

I am trying to use PROC SGPLOT with the following dataset. Currently with my script the x-axis goes from 1-52. I want the x-axis to start with week 35 and continue through week 52, then follow with weeks 1 to 34. Any assistance would be greatly appreciated. 

data original_data;
infile datalines dlm="," dsd;
input Surveillance_Week $ weekly_count_curryr rolling_avg_prevyr;
datalines;
35,15,7
36,11,5.7
37,,8.3
38,,9.7
39,,13.7
40,,15.3
41,,20
42,,24
43,,27
44,,27.7
45,,29.7
46,,32
47,,33
48,,30
49,,27.7
50,,24
51,,21
52,,15.3
53,,9.7
01,,13.7
02,,20
03,,18
04,,17
05,,14.3
06,,15.7
07,,14.7
08,,12.3
09,,10.7
10,,9.3
11,,11
12,,10.7
13,,10.3
14,,10.7
15,,11.3
16,,13.3
17,,13
18,,12.3
19,,9.7
20,,7.7
21,,7.3
22,,8.7
23,,7.3
24,,9
25,,8.7
26,,11.7
27,,10
28,,7.7
29,,5.3
30,,4.3
31,,6
32,,8.3
33,,7.3
34,,5.3
;
run;

proc sgplot data=comb_adjusted;
where serial_num >= 3;
vbar adjusted_week / response=weekly_count_curryr
datalabel
barwidth=0.8
fillattrs=(color=green);
vline adjusted_week / response=rolling_avg_prevyr
lineattrs=(color=blue thickness=3);
xaxis label="Surveillance Week"; /* No grid option here removes vertical lines */
yaxis label="count" grid; /* Keeps horizontal grid lines */
keylegend / position=topright across=1;
run;

[Ksharp · Accepted solution]

You'd better post a desired graphic to let me know what you are looking for .

data original_data;
infile datalines dlm="," dsd;
input Surveillance_Week $ weekly_count_curryr rolling_avg_prevyr;
datalines;
35,15,7
36,11,5.7
37,,8.3
38,,9.7
39,,13.7
40,,15.3
41,,20
42,,24
43,,27
44,,27.7
45,,29.7
46,,32
47,,33
48,,30
49,,27.7
50,,24
51,,21
52,,15.3
53,,9.7
01,,13.7
02,,20
03,,18
04,,17
05,,14.3
06,,15.7
07,,14.7
08,,12.3
09,,10.7
10,,9.3
11,,11
12,,10.7
13,,10.3
14,,10.7
15,,11.3
16,,13.3
17,,13
18,,12.3
19,,9.7
20,,7.7
21,,7.3
22,,8.7
23,,7.3
24,,9
25,,8.7
26,,11.7
27,,10
28,,7.7
29,,5.3
30,,4.3
31,,6
32,,8.3
33,,7.3
34,,5.3
;
run;

proc sgplot data=original_data;
vbar Surveillance_Week / response=weekly_count_curryr
datalabel
barwidth=0.8
fillattrs=(color=green);
vline Surveillance_Week / response=rolling_avg_prevyr
lineattrs=(color=blue thickness=3);
xaxis label="Surveillance Week" type=discrete discreteorder=data; /* No grid option here removes vertical lines */
yaxis label="count" grid; /* Keeps horizontal grid lines */
keylegend / position=topright across=1;
run;

 

[bijayadhikar]

Thank you Ksharp. This exactly what I was looking for. Much appreciated.

[ballardw]

---

@bijayadhikar wrote:

I am trying to use PROC SGPLOT with the following dataset. Currently with my script the x-axis goes from 1-52. I want the x-axis to start with week 35 and continue through week 52, then follow with weeks 1 to 34. Any assistance would be greatly appreciated. 

data original_data;
infile datalines dlm="," dsd;
input Surveillance_Week $ weekly_count_curryr rolling_avg_prevyr;
datalines;
35,15,7
36,11,5.7
37,,8.3
38,,9.7
39,,13.7
40,,15.3
41,,20
42,,24
43,,27
44,,27.7
45,,29.7
46,,32
47,,33
48,,30
49,,27.7
50,,24
51,,21
52,,15.3
53,,9.7
01,,13.7
02,,20
03,,18
04,,17
05,,14.3
06,,15.7
07,,14.7
08,,12.3
09,,10.7
10,,9.3
11,,11
12,,10.7
13,,10.3
14,,10.7
15,,11.3
16,,13.3
17,,13
18,,12.3
19,,9.7
20,,7.7
21,,7.3
22,,8.7
23,,7.3
24,,9
25,,8.7
26,,11.7
27,,10
28,,7.7
29,,5.3
30,,4.3
31,,6
32,,8.3
33,,7.3
34,,5.3
;
run;

proc sgplot data=comb_adjusted;
where serial_num >= 3;
vbar adjusted_week / response=weekly_count_curryr
datalabel
barwidth=0.8
fillattrs=(color=green);
vline adjusted_week / response=rolling_avg_prevyr
lineattrs=(color=blue thickness=3);
xaxis label="Surveillance Week"; /* No grid option here removes vertical lines */
yaxis label="count" grid; /* Keeps horizontal grid lines */
keylegend / position=topright across=1;
run;

---

Note that if you use a date value, such as the first day of the week, instead of an arbitrary number (1 - 52), that DATES will sort properly.

Something along these lines:

data original_data;
infile datalines dlm="," dsd;
input Surveillance_Week $ weekly_count_curryr rolling_avg_prevyr;
   if Surveillance_Week in (35:53) then weekdate = intnx('week','01JAN2022'd,Surveillance_Week-1);
   else weekdate =intnx('week','01JAN2023'd,Surveillance_Week-1);
   format weekdate date9.;
datalines;
35,15,7
36,11,5.7
37,,8.3
38,,9.7
39,,13.7
40,,15.3
41,,20
42,,24
43,,27
44,,27.7
45,,29.7
46,,32
47,,33
48,,30
49,,27.7
50,,24
51,,21
52,,15.3
53,,9.7
01,,13.7
02,,20
03,,18
04,,17
05,,14.3
06,,15.7
07,,14.7
08,,12.3
09,,10.7
10,,9.3
11,,11
12,,10.7
13,,10.3
14,,10.7
15,,11.3
16,,13.3
17,,13
18,,12.3
19,,9.7
20,,7.7
21,,7.3
22,,8.7
23,,7.3
24,,9
25,,8.7
26,,11.7
27,,10
28,,7.7
29,,5.3
30,,4.3
31,,6
32,,8.3
33,,7.3
34,,5.3
;
run;

proc sgplot data=original_data;
vbar weekdate / response=weekly_count_curryr
      datalabel
      barwidth=0.8
      fillattrs=(color=green)
;
vline weekdate / response=rolling_avg_prevyr
   lineattrs=(color=blue thickness=3);
   xaxis label="Surveillance Week" values= ('21AUG2022'd to '20AUG2023'd by week) 
               valuesrotate=vertical
   ;
   yaxis label="count" grid; /* Keeps horizontal grid lines */
   keylegend / position=topright across=1;
   format weekdate yyweeku3.;
run;

Note that it is not helpful to show a data set that does not match that used by the procedure. Your Original_data does not have have half the variables used in the plot which is using COMB_ADJUSTED

[bijayadhikar]

@ballardw  thanks very much for providing some useful tips. Noted. Correct, I did not include all vars thinking may not be needed. All is good.