Classify a raster with Method Defined Interval not possible

765
2
01-29-2022 04:35 AM
Heike_Janssen
New Contributor

I am working with ArcGIS Pro 2.8.4 but this problem occurred also in former versions (maybe since 2.6). I used to symbolise bathymetry rasters in defined intervals, e. g. 1 m. Since some months I am experiencing the problem that I cannot choose this method anymore in the symbology tab. ArcGIS switches automatically back to Natural Breaks. I can choose any method except the one I want/need. Works neither with .tif nor .asc files. I have tried the method in Arcmap 10.8.2 with an exact same raster file and it works fine. But I am not working anymore with ArcMap and really need this function in ArcGIS Pro.

Yes, I did calculate the statistics of the raster before trying to classify it. 

Any help is welcome. Thanks.

0 Kudos
2 Replies
DanPatterson
MVP Esteemed Contributor

Are the data floats? 

What is the range from min to max?

Data classification methods—ArcGIS Pro | Documentation

And you need to meet these criteria for it to work

The interval size must be small enough to fit the minimum number of classes allowed, which is three.

but if your range of values is too small or too large, then defined interval isn't appropriate which is why another method is chosen and you may need to provide a "manual interval"


... sort of retired...
0 Kudos
LenHealy
New Contributor

Were you able to solve this problem?

It seems that when all of the values of the grid are negative - or that the values default to a negative interval size when Defined Interval is selected then Pro won't allow switching to Defined Interval.

If you use the "Times" tool and multiply the raster by -1 it will change negative values to positive. You can import the symbology to original raster - then just change to Interval Size to a different number (and back if you need).

ArcMap has the same "error", but it lets you continue on to change the interval size - when the interval size is not negative, it allows the raster to draw as desired.

I don't know if this would be easy or not, but ESRI could address by having the interval size be an absolute value.

0 Kudos