topic Re: Pattern statement issues when device=javaimg in Graphics Programming

Pattern statement issues when device=javaimg

nbasnet — Thu, 26 Mar 2015 16:19:17 GMT

While using PROC GMAP and device=javaimg, the PATTERN statement functionality seems to work partially. For missing values, I have defined the color to be white and when I use device=png,gif,... , it seems to work fine. But when I use device=javaimg, the missing values color changes to other colors. Is there a trick to get around this?

Re: Pattern statement issues when device=javaimg

ballardw — Thu, 26 Mar 2015 16:54:06 GMT

Any chance you are using Options nogstyle? Javaimg doesn't honor that option.

Also are you using the CDEFAULT option if doing a choropleth map?

Re: Pattern statement issues when device=javaimg

nbasnet — Thu, 26 Mar 2015 17:04:10 GMT

No, I'm not using OPTIONS NOGSTYLE or CDEFAULT.

Here is the code I'm using:

pattern1 v=s c=white;

pattern2 v=s c=color2;

pattern3 v=s c=color3;

pattern4 v=s c=color4;

goptions hsize=12 vsize=9 device=javaimg;

Proc gmap data=zip map=map;

id zip;

choro rate/discrete missing

annotate=anno

coutline=black

legend=legend1

format rate g&form._as.;

run;

When I run the procedure, SAS ignores the pattern1 which is defined for missing values and puts a new color for missing values in the legends.

Re: Pattern statement issues when device=javaimg

Darrell_sas — Thu, 26 Mar 2015 17:14:00 GMT

Use CDEFAULT as an option to CHORO.

Re: Pattern statement issues when device=javaimg

nbasnet — Thu, 26 Mar 2015 17:18:34 GMT

CDEFAULT is not supported by java.

Re: Pattern statement issues when device=javaimg

ballardw — Thu, 26 Mar 2015 18:59:38 GMT

White should be assigned to the first non-missing value, formatted value, level or midpoint option. What value does your format g&form._as assign to missing?

Re: Pattern statement issues when device=javaimg

nbasnet — Thu, 26 Mar 2015 19:09:07 GMT

For missing, the value assigned is "." .

Re: Pattern statement issues when device=javaimg

ballardw — Thu, 26 Mar 2015 21:34:20 GMT

Since you don't appear to be requesting a statistic from the value I would try setting the missing to a value way outside your other data (-1e8 or similar) and include that value in your format.

With out seeing a map its hard to tell but I think your comment "missing values color changes to other colors" indicates that missing may be treated differently that an actual value.