Shift Tool + Parallel Processing Bug

695
1
04-04-2017 01:05 PM
AdamBorczyk
New Contributor III

I found an unexpected behavior occuring when using Shift Tool in ArcMap while having Parallel Processing environment value set to other than default (blank). It happens when I work on fresh Landsat data straight from Earth Explorer, previously untouched by any processing. I have described it briefly in this GeoNet thread, and below as well:

  1. having output location as default (that is default.gdb in your home folder) and Parallel Processing left blank: returns correct shifted image
  2. having output location as default (that is default.gdb in your home folder) and Parallel Processing set any proper value (eg. 50%) returns correct shifted image 
  3. having custom output location (inside one of Folder Connections) and Parallel Processing left blank: returns correct shifted image
  4. having custom output location (inside one of Folder Connections) and Parallel Processing set any proper value (eg. 50%) returns the same image - not shifted in any way. (Though other options will apply, like file format or cellzise) I tested it with different output file formats, none worked correctly.

This happens regardless of where do I set Environmental Values: tool's processing window or Menu->Geoprocessing->Environments.

While testing Shift Tool, I also gave Rotate tool a try - it acts similar, it doesn't work with Parallel Processing. I attached a prtscreen of input image for Rotate tool and an output of the tool (plus both together). You can see the black NoData area moved to a different place, but the actual data stayed there and was sliced at the top. I haven't tested other tools.

This problem was tested on ArcMap versions 10.4.5686 and fresh 10.5.0.6491 and different Landsat images. I think this is a bug that should be reported, please let me know what do you think.

Best regards,

Adam

0 Kudos
1 Reply
DanPatterson_Retired
MVP Esteemed Contributor

Since you have the data and the workflow, it might be a good idea to forward these data and examples through customer support so it can be investigated.

0 Kudos