I'm running Zonal Statistics as Table on three separate feature classes. These feature classes represent travel walk bands from various locations. These travel bands are my zone data, I have a regular raster stored in a geodatabase for my value raster, and I am calculating the mean raster value.

As you can imagine, many of these travel bands overlap, which is problematic for Zonal Statistics as Table. I got around this issue by using a Python script to iterate through each zone one at a time and then merge the outputted single-record tables at the end.

One outstanding issue I am having is that some of the input travel bands/zones refuse to calculate a mean raster value; it keeps returning a generic error. I have included a try-except clause in the script to handle these errors and include a note of the problem zones.

Isolating these problem zones and trying to run a Zonal Statistics as Table using the official tool returns the same error.

On a separate run of Zonal Statistics as Table, these same exact zones functioned perfectly fine with a different raster and using the sum statistic option.

I have tried various remedies including: ensuring projections match, checking field names, making there are not very many records, altering some of the suggested raster ArcMap options, repairing geometry etc...

The amount of records this error occurs with is quite small relative to the total number of zones, so I am not overly worried if I have to throw them out. Although naturally, I would like to include everything I can.

  • What are the size of the zones? Do they exceed the cell size of the raster or are they comparatively small? You mention the zones work with a different raster with sum statistic; do they work with this other raster and mean statistic? Are there any / a lot of null values in the problem raster in these zones? (could you try reclassifying?) – smiller Nov 16 '18 at 15:19