I've got a slight technical hitch that I hope someone may be able to shed some light on...
I have a shapefile layer consisting of several thousand polygons (Example1) alongside numerous raster datasets (Example 2). Essentially I need each polygon in the shapefile to take on the mean value of the cells it covers in the raster file.
What I've done so far is to use Zonal Statistics as Table and populate a new table with the mean values of the raster cells for each polygon using the shapefile as the feature layer. I then joined the new table with the original shapefile by matching the OID. This suceeded in creating exactly what I wanted, whereby each polygon has a mean value for each of the raster cells it overlays in it's attribute table. However, for some reason (that I find inexplicable) this method has arbitarily only worked for about half of the polygons that overlay the raster (Example 3), leaving a large proportion of the polygons with no value (as you can see by comparing the gaps in Example 3 with the raster dataset in Example 2). NOTE: these polygons DO cover underlying raster data, it is not a matter of checking/un-checking the Ignore NoData checkbox. I've repeated this method several times in both ArcGIS 9 and 10 and it still gives me the same results. I've also used model builder to iterate through the features in the shapefile and collate the results, but it still gives me exactly the same results as my initial, simpler method (Example 3).
Some help/suggestions as to why this is occurring would be brilliant as it's a rather essential part of the project.
