4

I have to perform ArcGIS zonal statistics on a raster, but the column I am interested in is not the 'value' column. Now, I can reclass the raster so that the column of interest becomes the 'value' column. However, the reclassification is super slow (reclassifying a 60 m raster for the US Midwest, length of reclass ASCII file ~500,000 rows.)

Is there any way I can skip the reclassification and just perform zonal statistics on my column of interest? Alternatively, any tips on making the reclassification process much faster will help too. Currently the reclassification is on course to take 10 hours to finish.

2
  • How are you doing the reclassification? Commented Feb 24, 2014 at 22:34
  • Using reclassbyascii file command in ArcGIS desktop.
    – user1186
    Commented Feb 24, 2014 at 22:44

1 Answer 1

1

If the values in your value field correspond with unique values in your field of interest you could run zonal statistics on the value field, then join the output with the original table and then calculate the value of your field of interest.

I've done this to populate a model grid with soils data. Each grid cell is given a unique number - I calculated this from the OBJECTID field (I don't think you can use this field as your "zones" field). The soils data consisted of a 100m grid with 7 different hydrologic soil group codes:

hydrologic soil group code raster attribute table

I ran zonal statistics using my grid id as the zone field and the value field of my raster to return the majority statistic to a table:

enter image description here

I joined the table back to the original raster to grab the "hydrgrpdcd" value I wanted:

enter image description here

From there I did a join on the zonal statistics output table with the model grid feature class using the "grid_id" field and assigned the hydrologic soils code.

I did this for roughly 500,000 records and it took just a few minutes to compute the values of the new field.

There are probably better ways to do this, but this worked for me. Hopefully this makes sense!

1
  • 1
    Thank you so much Jason, unfortunately, my field of interest has multiple values that are the same preventing me from using this approach.
    – user1186
    Commented Feb 24, 2014 at 21:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.