topic Re: Gchart - Is it a bug while sorting data by descending values to a gchart? If that, how to fix it? in Graphics Programming

Gchart - Is it a bug while sorting data by descending values to a gchart? If that, how to fix it?

tanyuanbin — Mon, 18 Mar 2013 07:28:51 GMT

Dear All,
Can I bother you second, it seems that Proc Gchart might produce wrong output if we choose to sort by descending values.
1. Correct output by ascending value as below:

2. Incorrect output by descending values as below:

The source code as such:

%let Stock_cd = IF1304;

ods listing;
filename grafout "C:\Temp\jbs_desc.gif";

goptions reset=all
          device=gif
          gsfname=grafout
          gsfmode=replace;

GOPTIONS xpixels=400 ypixels=300;

GOPTIONS COLORS = (BLUE, BLUE, CYAN, CXD2477C, Green, Red);

Legend1
FRAME
POSITION = (BOTTOM CENTER OUTSIDE);

Axis1
     STYLE=1
     WIDTH=1
LABEL=( "Current Price" )
;
Axis2
     STYLE=1
     WIDTH=1
     MINOR=
     (NUMBER=1
     )
;
TITLE;

TITLE1 h=2 "&Stock_cd. Price Distribution Analysis" -2D. ;

FOOTNOTE;

PROC GCHART DATA=Interest.Sample2 gout=work.gseg;
     HBAR     Current_Price
/ SUMVAR=Trans_Amount
     CLIPREF
     DESCENDING

FRAME SUM /* NOSTATS */
     TYPE=SUM
     LEGEND=LEGEND1
     COUTLINE=BLACK
     MAXIS=AXIS1
     RAXIS=AXIS2
     LREF=34
     CREF=BLUE
/*     AUTOREF */
CTEXT=BLACK
CAXIS=BLACK
;

RUN; QUIT;

Thank you in advance for your time!

Best Regards,

Yuanbin

- and part of those raw data:

Current_Price group Trans_Amount

2584 B 78

2593.4 B 130

2594.6 B 26

2594.6

2594.6 B 26

2594.6 -26

2593.2 26

2592.8 52

2592.2 S -26

2592.8

2592

2592.8 B 26

2592.8

2593.2

2593.6 B 26

2592.8 S -26

2593

2593 B 26

2592.8

2594

2593.8

2593.8 S -26

2593.8

2594.4

2594.4 -52

2603.2 S -26

2604.6

2603.8 26

2604.4 -26

2603.8

2603.4

2604.6 -26

2604.6 S -26

2603.8 S -26

2603.4 26

2603

2604

2603.2

2602.8

2601.8 B 26

2602.2 -26

2601.8

2601.2 B 26

2601 52

2600.8 S -26

2689.6 27

2690.2 -27

2690.2 27

2690.6

2690.6 S -27

2691.2 -27

2691.2 27

2691

2691.8

2692

2691.8

2691.8 S -27

2691.8

2691.4 -27

2690.8

2691.4 -54

2691.6 27

2692

2691.6 377

2691.2

2691.2 27

Re: Gchart - Is it a bug while sorting data by descending values to a gchart? If that, how to fix it?

LinusH — Mon, 18 Mar 2013 09:52:30 GMT

The ASCENDING/DESCENDING options to the HBAR (and other GCHART statements) specifies a sort order based on the statical values/measures, so no, it's not a bug.

Ordering for midpoint/class variables can be specified on the AXIS statements.

Re: Gchart - Is it a bug while sorting data by descending values to a gchart? If that, how to fix it?

tanyuanbin — Mon, 18 Mar 2013 14:27:54 GMT

Hi LinusH,

I appreicated it very much for your time, it really help!

after I add option of ORDER=(2700 to 2580 by -4 ), the chart looks good:

it's wonderful! by the way, do you know how to apply for a default value for the step-length in the value-list? for example here as of 'by -4', otherwise, I have to define them in source code, If I ignore it, for example, 'ORDER=(2700 to 2580 ), it will order by ascending again?! Do you have any advice, or would you help to show me more in details? Thanks in advance!

Re: Gchart - Is it a bug while sorting data by descending values to a gchart? If that, how to fix it?

LinusH — Mon, 18 Mar 2013 14:41:43 GMT

It was some time since I did GCHART...

I don't think there is a default value for stepping. If you omit it, does your ?! it does ascending...?

You could have macro variable with your default stepping, and always specify it in your AXIS...

May not help as much, I think you be better off dig into the documentation and test different approaches yourself...

Re: Gchart - Is it a bug while sorting data by descending values to a gchart? If that, how to fix it?

tanyuanbin — Tue, 19 Mar 2013 01:52:12 GMT

After took more look into the output of gchart, it seems there is an automatical stepping algorithm which displays with different step size, as a result the chart of bar looks better and continuously display as well.

Thanks a lot for the userful information!

Yuanbin