Raster Processing - 010092, Invalid output extent: It Persists.

2661
4
08-24-2017 06:38 AM
TylerSearls2
New Contributor III

Hoping to resolve a persistent error I've been receiving when trying to employ raster processes in Python: 010092, Invalid output extent. I'm projecting with Canada Lambert Conformal Conic. I'm fairly confident the syntax is correct; given the fact both the Euclidean Distance and Line Density processes fail with the same input layer, I'm suspicious of the input.

I should add that when I run the Euclidean Distance or Line Density processes directly through ArcGIS Pro the processes (seem to) run using the same dataset called as the input in the script. I haven't run the processes all the way through in ArcGIS Pro to confirm this as it's a national dataset and would take substantial time, but it does (start to) run. Fresh perspectives are appreciated.

0 Kudos
4 Replies
DanPatterson_Retired
MVP Emeritus

Being that someone... it may be time to show the calling script to see what might be causing the error

The error means...

Description

The extent specified, either in the environment or as a parameter, is invalid for the output that will be created.

TylerSearls2
New Contributor III

Hello again Dan_Patterson‌, thanks for your continued assistance, the python file has been posted.

0 Kudos
DanPatterson_Retired
MVP Emeritus

Tyler

There is no guarantee that the dataframe nor the other files are in the Lambert projection.  It will bounce back and forth between mixed coordinate systems.  I would strongly suggest that you separate the project stuff out from the rest.  

Also when converting some of the files to raster, you need to explicitly set the environment properties to those of the main raster.  It should serve to set the cell size, the extent and it should also be the snap raster

To this end, examine the arcpy topics here especially the env environment property which has read/write properties and the Extent property (read) You should be checking the extent object.  

If it were me, I would project all the data to Lambert and be done with it, then be careful to set the cell size, extent and snap raster before running your script. Also, dump anything out of the dataframe that isn't part of the analysis ie basemaps and anything unnecessary... check your coordinate systems of the data and the data frame and go from there

curtvprice
MVP Esteemed Contributor

I'm totally with Dan on this, with 25 years of experience with Spatial Analyst / Grid. What really kills you in asking ArcGIS to project data on the fly is the extent setting. Extents are straight line boxes and do not project well, even if the software attempts it. Also think about cell size - imagine specifying a 300m cell size and your workflow is set up in your script to have that be 300 'units' which get handled with geographic as 300 degrees -- you get the idea? Projecting ahead of time is much safer.