topic Re: Zonal Statistics as Table Not Working in ArcGIS Spatial Analyst Questions

Zonal Statistics as Table Not Working

KateBallard — Mon, 17 Jul 2017 22:57:54 GMT

I am trying to run the Zonal Statistics as Table tool on my data in order to sum the number of cells of a certain type of landcover in each county. For example, I want it to sum how many Deciduous forest, Pasture/Hay land and etc.. there are in each county individually.

I have tried two different ways: 1. By using the Extract by Attributes tool on the raster to extract each individual type of cell (so, one raster is created for deciduous forest cells and one raster for paster/hay and etc...), then running the zonal statistic on each of those one at a time (not sure if this is necessary or not). 2. By just running the Zonal statistics as Table tool on the full raster.

I have tried EVERYTHING (saving it in a personal geodatabase, saving it in a file, getting rid of spaces in file names, making sure zone feature name is less than 17 characters etc...) and I still keep getting either "ERROR 999999" or there is no error at all under messages or even an indication that it has stopped, instead the red X circle just appears next to

it with no message. I attached a screenshot of what that second occurrence looks like.

Does anyone have any idea what I might be able to do to make this tool work? And do I need to be performing the "Extract by Attributes" tool or should the Zonal Statistics tool be able to sum up each type of cell (e.i. deciduous forest, developed land etc..) individually in each county?

Re: Zonal Statistics as Table Not Working

DanPatterson_Retired — Tue, 18 Jul 2017 00:23:03 GMT

from your screen grab...

there appear to have spaces in the file names and (Tiger....)

what is the field with the period in it? it suggests a join on the table (not good)

input value raster (no file extension... what is it? not an esri grid for sure

Z folder... is that a network drive? a remapped drive? not a usb or some other non-local drive

Did you try a small sample of the data (split some pieces out to rule out memory issues)

Re: Zonal Statistics as Table Not Working

JayantaPoddar — Tue, 18 Jul 2017 05:28:02 GMT

Disable background processing and try again.

Re: Zonal Statistics as Table Not Working

KateBallard — Tue, 18 Jul 2017 15:34:24 GMT

- I went back and changed the file names to have "_" instead of spaces, the one you see is just me having edited the name in properties to be more readable.

- Yes, I joined tables together to match my Lyme disease case numbers with the county shapefiles, is that a problem for trying to perform a zonal statistic? If so, how should I go about fixing this and getting the Lyme case data and population data matched with the shapefile?

- No its not from esri, the input value raster was downloaded from the National Land Cover Database (NLCD).

- Yes, the Z folder is a server that I am doing all of this work on because I am working on a university computer, is that a problem for this tool? I can move it to the computer and then move it back to the server before I shut down the computer. Will that cause any issues or just involve restoring the pathnames?

- I tried the normal (non-table) Zonal Statistics tool on one county, and it just gave me a black box over the area. I assume that means the tool didn't work. I also tried the Zonal table tool on just one of the cell types, and that also didn't work.

Thank you for all your help!

Re: Zonal Statistics as Table Not Working

ChrisDonohue__GISP — Tue, 18 Jul 2017 17:07:28 GMT

Dan Patterson and Jayanta Poddar hit the highlights of what can go awry with this sort of process.

Setting up the data is crucial and often requires a different method of file naming and structuring than with most other types of GIS processing. Oftentimes this will involve unique data and folder set up before running the process to help ensure the processing will work:

No spaces or restricted characters in the file names nor the folder names. So no spaces, no parenthesis. For example, the input file TIGER County Boundaries (2015) will need the spaces and parenthesis removed.
Keep the path to the data short. While technically one can put the data out several folders in a path, in reality this can cause processes to crash.
Keep filenames very short. As a general rule, I'd shoot for 8 characters or less. They can always be renamed later.
Avoid using network drives if possible. Your hard drive is the best way to go if you have access to it and available space.
Try putting the data into a simple folder, like C:\Data
In regards to the Joined data, save off the Joined data to a new feature class (Data, Export Data) and run the process on that new feature class.
"Disable Background processing" often is the cure when otherwise everything else seems set up OK.

Chris Donohue, GISP

Re: Zonal Statistics as Table Not Working

KateBallard — Tue, 18 Jul 2017 21:45:47 GMT

Thank you all for the help! It's working now. It seems the length of the filenames and the joined table were the problems.

It appears that I cannot run the Zonal Statistics as table tool and have it sum up each cell type individually (for example number of deciduous cells in each county, number of coniferous cells etc in each county). Instead it just seems to sum all the cells together.

The only fix I know for this is to run Extract by Attributes for each cell type (which takes a very long time), then performing the Zonal Statistics tool on each one. Does anyone know of a tool that could allow me to do this in a more efficient way, or to select that I want a sum of each type of cell individually?

Thank you!

Re: Zonal Statistics as Table Not Working

DanPatterson_Retired — Tue, 18 Jul 2017 22:03:58 GMT

Did you try to make the file with the join permanent by exporting to a new featureclass, then copy it to a local folder with a simple path (ie c:\test) You have to simplify the process and joined data, complex paths, networks and files that are too big just complicate issues. Give it a go

Re: Zonal Statistics as Table Not Working

ChrisDonohue__GISP — Tue, 18 Jul 2017 22:40:48 GMT

To gather the Zonal Statistics on a County-by-County basis one will likely have to employ either Modelbuilder or Python to loop through each county and do a Zonal Statistics run on just that county.

For Modelbuilder, the looping would be via an Iterator

Note that one can carry over some of the information to be used in the file naming via the use of In-line Variable Substitution. For example, with this process as each county is run the County name (or other information) could be added as part of the output file name for each result to make each output more easily recognizable.

A quick tour of using inline variable substitution—Help | ArcGIS for Desktop

If you are familiar with Python, there are a variety of looping methods.

Chris Donohue, GISP

Re: Zonal Statistics as Table Not Working

KateBallard — Tue, 18 Jul 2017 23:23:51 GMT

Yes I did exactly that, shortening file names and making the join permanent by exporting it to a new feature class, except I kept it on the outside server in the same folder as the main map, and the tool luckily still worked. Thank you!

Re: Zonal Statistics as Table Not Working

DanPatterson_Retired — Tue, 18 Jul 2017 23:33:22 GMT

Good... You might want to close the thread Kate, before you get any more advice good luck with your studies

Re: Zonal Statistics as Table Not Working

KateBallard — Tue, 18 Jul 2017 23:34:46 GMT

Sorry, I don't think I explained very well. I actually got the tool to work (summing for each individual county) by just associating it with a code in the shapefile that was unique to each county. My issue is that in the raster there are different types of cells (blue cells = water (11), dark green cells = evergreen forests (42) and etc). I basically am using the Zonal Statistics as Table tool to tell me how many water cells there are in each county, how many evergreen forest cells in each county and so on. So far, the only way that I can find to have it sum water cells and evergreen forest individually, rather than just summing all cells in each county, is to first use "Extract by Attributes" to make an individual raster for each landcover type (so a raster showing only water cells, a raster showing only evergreen forest cells, etc..) Then running the Zonal Stat as table tool on each one of these separately. I'm wondering if there is an add-on or other tool that would allow me to tell the Zonal statistic tool to sum by cell type? Unfortunately I'm not trained in Python.

My other question is, I want to tell the Zonal Stat tool to only count cells that are completely inside a county, not ones that are only half inside a county (like the cell is split by the county boundary so only half the cell falls in the county). However, I do not see an option for this in the Zonal Stat tool window. Does anyone know a way that I could accomplish this?

Thank you!

Re: Zonal Statistics as Table Not Working

KateBallard — Tue, 18 Jul 2017 23:40:24 GMT

I will do that! I have one more question if you don't mind, I want to tell the Zonal Stat tool to only count cells that are completely inside a county, not ones that are only half inside a county (in other words exclude raster cells in the SUM that are split by the county boundary line). However, I do not see an option for this in the Zonal Stat tool window. Do you have any idea how I could accomplish this?

Re: Zonal Statistics as Table Not Working

DanPatterson_Retired — Tue, 18 Jul 2017 23:49:49 GMT

That isn't how a raster cell works... the center of the cell is inside... or not... don't get carried off by the visual representation of the square, it is the centre that is used.

Re: Zonal Statistics as Table Not Working

XanderBakker — Tue, 18 Jul 2017 23:58:12 GMT

From the Help:

If the zone input is a feature dataset, a vector-to-raster conversion will be internally applied to it.

The tool will not allow you to control it. As an alternative you could convert the zones to raster first using Polygon to Raster—Help | ArcGIS Desktop. This allow some control over the cell assignment, but not that it will only use the cells that are completely inside a polygon. You could, convert the raster representation to vector again, do a spatial select and work on that, but please understand that your statistics will be off since pixels that are shared by two zones will not be taking into account.

Re: Zonal Statistics as Table Not Working

KateBallard — Wed, 19 Jul 2017 01:23:02 GMT

Oh I see. So, that means that a cell "split" by a county line will only be counted in the sum of one of the counties and not both, because the center can only fall inside one county (zone)?

Re: Zonal Statistics as Table Not Working

XanderBakker — Wed, 19 Jul 2017 01:30:34 GMT

That is correct! Since the zones are converted to raster first, a pixel can only be part of a single zone. If you set the environment setting to match the cellsize and snap raster to the raster the you want to extract the statistics from, you will have no problems. Convert the zones to raster first if you want to have a better control of how a pixel is assigned to a zone.

Re: Zonal Statistics as Table Not Working

DanPatterson_Retired — Wed, 19 Jul 2017 01:32:09 GMT

That is correct. Converting a polygon to a raster is going to show this behaviour if you zoom in close enough, and as Xander points out, you can gain control over the cell size of the conversion by doing the conversion prior to performing any further analysis. Processing Environments should always be considered and set before doing any sort of analysis. Unfortunately the 'Environments' button is way at the bottom right of all the tools in Arctoolbox in the region 'obscurata'... at least in ArcGIS Pro.. Environments is placed on an equal footing on the top of all tool dialog. Never assume... especially when working with rasters... by default the cell size is 1/250th the largest of the extent (width or height)... but what if you want a cell size of 10 m? isn't going to happen unless you set it before doing your work.