You are here: > ESRI Forums > arcgis desktop discussion forums > Thread Replies

ArcGIS Desktop Discussion Forums

ArcGIS Desktop - Raster display and management forum

Raster calculator "error"   Vic Bogosian May 28, 2009
Re: Raster calculator "error"   William Huber May 29, 2009
Re: Raster calculator "error"   Vic Bogosian May 29, 2009
Re: Raster calculator "error"   Vic Bogosian May 29, 2009
Re: Raster calculator "error"   Vic Bogosian May 29, 2009
Re: Raster calculator "error"   William Huber May 30, 2009
Re: Raster calculator "error"   Vic Bogosian May 30, 2009
Re: Raster calculator "error"   Raychl Lukie Jul 05, 2009
Re: Raster calculator "error"   William Huber Jul 05, 2009
Re: Raster calculator "error"   Raychl Lukie Jul 05, 2009
Re: Raster calculator "error"   William Huber Jul 06, 2009
Report Inappropriate Content • Top • Print • This Forum is closed for replies.    
Subject Raster calculator "error" 
Author Vic Bogosian 
Date May 28, 2009 
Message I'm creating raster layers via calculator using ecological distance statistics. The final output can range from zero to infinity. My initial attempt at creating a final output map did not produce any visible results (calculated layer and permanent raster from it exist, take up ~380 MB of space, but do not display anything or have any pixels I can detect). My output value range is from 1.1e-38 - 3.4e+38, which looks like it may be too much for the available raster pixel depths (32 bit, etc) to handle. What are my options for displaying these kinds of data?

Thanks,

Vic 
   
Report Inappropriate Content • Top • Print • This Forum is closed for replies.    
Subject Re: Raster calculator "error" 
Author William Huber 
Date May 29, 2009 
Message Re-express the results, Vic. A good choice for positive values like yours is the logarithm. The presence of these extreme values indicates both floating point underflow and overflow occurred during your calculation, so you will likely need to redo it in a different way (perhaps by introducing logarithms at an earlier stage of the process). 
  --Bill Huber
Quantitative Decisions (http://www.quantdec.com )
More GIS Q&A at http://gis.stackexchange.com/q/3083/664 
   
Report Inappropriate Content • Top • Print • This Forum is closed for replies.    
Subject Re: Raster calculator "error" 
Author Vic Bogosian 
Date May 29, 2009 
Message That seems to help, but I have run into another issue. Some rasters have nodata values in locations where other rasters actually have data, but my output seems to be limited strictly to pixels that all contain values (i.e. any pixel in any raster that contains nodata nullifies any output). Besides reclassifying each raster, is there a step around this in raster calculator?

Thanks,

Vic 
   
Report Inappropriate Content • Top • Print • This Forum is closed for replies.    
Subject Re: Raster calculator "error" 
Author Vic Bogosian 
Date May 29, 2009 
Message I should add that a bit of searching on the help forums has revealed a way to change nodata values to zero, but what I need is to change them to a floating point decimal of my choosing (specific to each raster). I can do this in a seperate Python script, but my datasets contain 100 million+ pixels each, and if there is an easy way to do it in ArcMap I would rather go with that route.

What I need is either a reclassify function that will only alter nodata values to a float and leave everything else alone, or a raster calculator operation that will do the same. Or a faster computer to run several 380 MB files through a Python script!

Thanks again for any help,

Vic 
   
Report Inappropriate Content • Top • Print • This Forum is closed for replies.    
Subject Re: Raster calculator "error" 
Author Vic Bogosian 
Date May 29, 2009 
Message I believe I have found a solution to my own problem (for once, so it may be time to buy some lottery tickets). This is a fairly specific-sounding problem, but if anyone else is experiencing it here is how I solved it.

If you run IsNull([MyGrid]), you will generate a map of 0 and 1 values for your nodata/data containing pixels. You can then reclassify/use raster calculator to give the nodata cells the value you want to (eventually) be your floating data points by making it 1000 times bigger (i.e. if you want it to be 0.483, make nodata = 483) and then dividing it in calculator by 1000 (the reclassify function only lets you create integer new values). The final step is to mosaic together your original and your new raster and choose the maximum overlay rule (your final calculation will be 0.483, which will overwrite any nodata cells). This is fairly convoluted and requires making several temporary datasets per raster, and I expect a simpler solution exists.

Thanks for the help,

Vic 
   
Report Inappropriate Content • Top • Print • This Forum is closed for replies.    
Subject Re: Raster calculator "error" 
Author William Huber 
Date May 30, 2009 
Message It's easier, faster, and more accurate to use the conditional operator, Vic, as in

Con( IsNull([MyGrid]), 0.483, [MyGrid] )

You are correct that this has to be done for each individual input grid. It can be done on the fly, however. For example, to find the maximum of two overlapping grids (treating NoData as equal to 0.483), you can compute

Max( Con( IsNull([Grid1]), 0.483, [Grid1] ), Con( IsNull([Grid2]), 0.483, [Grid2] )) 
  --Bill Huber
Quantitative Decisions (http://www.quantdec.com )
More GIS Q&A at http://gis.stackexchange.com/q/3083/664 
   
Report Inappropriate Content • Top • Print • This Forum is closed for replies.    
Subject Re: Raster calculator "error" 
Author Vic Bogosian 
Date May 30, 2009 
Message Bill -

Thanks, that is much simpler. Thanks again for your help.

Vic 
   
Report Inappropriate Content • Top • Print • This Forum is closed for replies.    
Subject Re: Raster calculator "error" 
Author Raychl Lukie 
Date Jul 05, 2009 
Message I have tried this calculation but I still get the same results.
I have three raster layers that I am adding together and when I add them them the summation layer only covers up to the area for the smallest raster area and not beyond this layer, despite the fact that I have 2 other raster layers that extend further. Anyone have any other ideas?
Thanks, Raychl 
   
Report Inappropriate Content • Top • Print • This Forum is closed for replies.    
Subject Re: Raster calculator "error" 
Author William Huber 
Date Jul 05, 2009 
Message Set the output extent in the Spatial Analyst "options" dialog or in the ArcToolbox "environment" dialog (depending on how you're doing the summation), Raychl. 
  --Bill Huber
Quantitative Decisions (http://www.quantdec.com )
More GIS Q&A at http://gis.stackexchange.com/q/3083/664 
   
Report Inappropriate Content • Top • Print • This Forum is closed for replies.    
Subject Re: Raster calculator "error" 
Author Raychl Lukie 
Date Jul 05, 2009 
Message I have tried adjusting my environmental settings under ArcToolbox, my settings under tools>options, and also my settings under spatial analyst and I keep getting the same results. Any further suggestions? I wonder if my raster layers are too large so I will try a smaller subset, but please let me know if you have something else I can try in the meantime.
Thanks, Raychl 
   
Report Inappropriate Content • Top • Print • This Forum is closed for replies.    
Subject Re: Raster calculator "error" 
Author William Huber 
Date Jul 06, 2009 
Message You need to provide some details, Raychl. Consider starting a new thread rather than piggy-backing on this one. Describe exactly what you have for data, what you're doing with them, and what the output is. 
  --Bill Huber
Quantitative Decisions (http://www.quantdec.com )
More GIS Q&A at http://gis.stackexchange.com/q/3083/664