cloning environment best practices

[clt_cabq]

I installed Pro 3.4 and of course I have to re-clone my python environment, and i was wondering if there is a 'best practice' for doing this - is it best to create the environment in the same directory structure as the built-in environment, or can/should it be created elsewhere? Does any of this change when working with VSCODE or other IDE? 

[DanPatterson]

If it is within the existing structure, it makes it easier to maintain/upgrade your clones when you upgrade pro.

There are existing documents that may be appropriate on this topic

Clone an environment—ArcGIS Pro | Documentation

Package Manager—ArcGIS Pro | Documentation

the latter has information regarding network access amongst other things.

[clt_cabq]

Thanks Dan, I've read through a couple of your blogposts and some of those other threads. I think at one point you have commented you didn't like the original install location of the python environment so preferred to do it elsewhere with a less complicated path, but I think that was for earlier versions. 

[DanPatterson]

I install Pro in a simple path

c:\arc_pro

It is really easy to find in windows file explorer

arcpy and arctoolbox code can be found in

C:\arc_pro\Resources

python and the environments are in the path

C:\arc_pro\bin\Python

and the actual pro environment is in

C:\arc_pro\bin\Python\envs

    C:\arc_pro\bin\Python\envs\arcgispro-py3

    C:\arc_pro\bin\Python\envs\...your_clone_here

which is where I put my clones

Site packages are located in

C:\arc_pro\bin\Python\envs\arcgispro-py3\Lib\site-packages

Of course a "clone" copies over all the bloat that is in there, so I normally create an environment in the ...env path using conda and install the packages that I need there.  Usually my environments don't include arcpy since I am working with specific packages that are not compatible with arcpy (eg. numpy 2.x) which can be installed using specific conda commands and it keeps the other environment separate from the main one needed for pro and arcpy.