mkoohafkan / arcpy Goto Github PK

should use collapse =, not sep =

da_read(layer)

Loading required namespace: tibble
Error in as_tibble(x, stringsAsFactors = FALSE) :
could not find function "as_tibble"

There are some issues with directly using the environment as cloned from within the ArcGIS Pro application. No idea why but something about the way ArcGIS Pro clones the environment doesn't play nice with reticulate. It appears that there are two solutions:

1. After cloning the environment within ArcGIS Pro, clone it again from the anaconda prompt and then use that environment (the base arcpy environment doesn't show up in the environment listing, but maye you can clone it directly by providing a full path).
2. Following this ESRI documentation, create an environment from scratch and install arcpy and/or argis modules directly from the ESRI conda channel. One issue with this is that you need to install the module version from the channel that matches with your ArcGIS Pro installation, i.e., conda install --name <conda env> -c esri arpy=<Pro Version>

Preliminary tests suggest that option 1 works.

d = da_read(layer, "SHAPE@")

Error: Specified fields do not exist in table: SHAPE@

list of shape tokens: http://pro.arcgis.com/en/pro-app/arcpy/get-started/reading-geometries.htm

But there are more than this, e.g. OID@.

Maybe a way to check if a token is valid for a given layer? Or maybe just ignore checks for fields that have @ symbol.

conda search arcpy -c esri provides a listing of arcpy module versions:

# Name                       Version           Build  Channel
arcpy                    2.3.1.15800 py36_arcgispro_1  esri
arcpy                            2.7 py37_arcgispro_26794  esri
arcpy                            2.7 py37_arcgispro_26810  esri
arcpy                            2.8 py37_arcgispro_29717  esri
arcpy                            2.8 py37_arcgispro_29724  esri
arcpy                            2.8 py37_arcgispro_29734  esri
arcpy                            2.9 py37_arcgispro_32649  esri
arcpy                            2.9 py37_arcgispro_32704  esri
arcpy                            2.9 py37_arcgispro_32739  esri
arcpy                            3.0 py39_arcgispro_36045  esri
arcpy                            3.0 py39_arcgispro_36056  esri
arcpy                            3.0 py39_arcgispro_36057  esri
arcpy                            3.1 py39_arcgispro_41748  esri
arcpy                            3.1 py39_arcgispro_41751  esri
arcpy                            3.1 py39_arcgispro_41759  esri
arcpy                            3.1 py39_arcgispro_41791  esri

Output can be retrieved programmatically via flag --json. Given the ArcGIS Pro version (supplied by the user) it is therefore possible to determine the required python version automatically rather than having the user provide it.

Retrieving the module list can be implemented right now via

jsonlite::fromJSON(system2(reticulate::conda_binary(), c("search", "arcpy", "-c", "esri", "--json") stdout = TRUE)

but if reticulate implements conda_search() it would be a little cleaner.

For vector layers:

• https://pro.arcgis.com/en/pro-app/latest/tool-reference/conversion/features-to-json.htm
• https://pro.arcgis.com/en/pro-app/latest/tool-reference/conversion/json-to-features.htm

For rasters:

• https://pro.arcgis.com/en/pro-app/3.1/arcpy/functions/rastertonumpyarray-function.htm
• https://pro.arcgis.com/en/pro-app/latest/arcpy/functions/numpyarraytoraster-function.htm