Latest Contributions by AndersMark

When using Oracle you can use TNS and create a generic name for a datasource. This name can then be used from ArcGIS Database connection. That means that the exact same service can be published on several Arcgis servers and by configuration use different databases. There seems to be a similar way of doing this in Postgres called something like PGSERVICE. A layer between the caller (psql, Arcgis) and the actual server and database.  I have made a few tests and got this to work when using Postgres sql tool, called psql. Now the question is: How do I make a connection from ArcGIS and use the configuration in PGSERVICEFILE ?   Example: I have a Postgres database set up on another machine. From ArcGIS PRO I want to connect to that database by using definitions in the connection service file (https://www.postgresql.org/docs/current/libpq-pgservice.html) If we take the example from page above I would like to make a database connection i AGP that uses the [mydb] entry.  [mydb] host=somehost port=5433 user=someuser password=somepassword I have made adjustments in my local service file and can verify that it works by making a call to"psql", e.g: "C:\Program Files\PostgreSQL\14\bin\psql.exe" -d service=mydb This login works perfect. Now I want to use "mydb" from database connections but cant figure out the syntax. Anyone who has tried this? What I would like to recreate here is the same behaviour as you can with Oracle and where you can make a database connection and just use the TNS-names.    ... View more

I'm trying to find an error which occurs when running an FME-script which calls some Python code. In order to recreate the problem I use ArcMap to execute Python code and there I discovered some strange things.  Arcmap version 10.7 First I use toolbox to execute the method: arcpy.MakeRasterLayer_management I enter all values which in this case is just an input and output raster. The call works well. From the Results section I copy the Python Snippet which gives me the call in Python.  Then I paste this code into the ArcMAP Python Console and when executing i get this error: Runtime error Traceback (most recent call last): File "<string>", line 1, in <module> File "c:\program files (x86)\arcgis\desktop10.7\arcpy\arcpy\management.py", line 7519, in MakeRasterLayer raise e ExecuteError: ERROR 000867: Environment: contains invalid cell size or dataset. The call looks like:  arcpy.MakeRasterLayer_management(in_raster="D:/tmp/anders/Markatkomst_NMD.gdb/NVA_raster", out_rasterlayer="in_memory\NVA_raster2_mem", where_clause="", envelope="281360 6153230 820480 7403330", band_index="") In short I'm trying to copy a raster from a file gdb to in_memory for later use.  Anyone who has an idea what the problem is? Is this a bug?   My original code in ArcMap which doesn't work either looks this simple:  import arcpy import os workSpace = 'D:\\tmp\\anders\\Markatkomst_NMD.gdb' arcpy.CheckOutExtension("Spatial") arcpy.env.workspace = workSpace arcpy.env.extent = "MINOF" arcpy.env.overwriteOutput = True raster_name = "NVA_raster" raster_mem = "in_memory\\" + raster_name raster = os.path.join(workSpace, raster_name) print('Make raster layer prio') print(raster) arcpy.MakeRasterLayer_management(raster, raster_mem)   Regards Anders Mark         ... View more

I've run into a problem with an Oracle View that produces duplicates of column OBJECTID. The underlying geo-table has of course unique objectid:s but in the view, every geometry can occur severeal times. This violates functionality that requires that the ObjectID be unique. https://desktop.arcgis.com/en/arcmap/latest/manage-data/using-sql-with-gdbs/object-id.htm The view is a registred view in ArcGIS.   In my case I have a ArcMap service that consumes the view and when duplicate objectids are spotted it aggregates all rows with the same objectid. Maybe thats fine with the geometrys, as many lines can exist in a poly line field. The problem is the attributes where only one of aggregated rows will represent the geometry/row when consumed in the map service. Maybe it's the first row of the aggregated records that will display in the map or one of the others, it's hard to tell. It feels random.  Is there a best practise to solve this kind of issue? I have tried to make the view generate it's own unique value in the OBJECTID column instead reading that's  written in the underlying table but when I do this it's not possible to register the view in ArcGIS.    ... View more

I got a strange problem that occur when I combine rasters in my pythoncode.  Say I have rasters "NMD_COMBINE", "DosAreaRaster_r" and the result ends up in "rasterOut". Call in code: arcpy.gp.Combine_sa(["NMD_COMBINE", "DosAreaRaster_r"], "combineRaster") Actually all these rasters are "in_memory" so the actual path for all of them are : "in_memory\\NMD_COMBINE" "in_memory\\DosAreaRaster_r" "in_memory\\combineRaster" The call to combine is fine but the result is a little strange regarding the given column names in the "combineRaster".  The first 3 column are standard: "OID", "Value", "Count". The other 2 for the combined rasters are: "NMD_COMBIN" "gptmpraste" Look at the columns above! The first one has almost a correct name but are truncated to 10 characters. The second has a completely different name compared to the source but is also 10 chars long. Doesn't it look like it's also truncated and that an "r" is missing?  Question: Why does this happen? Normally this Pythoncode is embedded in FME-script but is now extracted and run i Arcmap 10.3 Python window.  If I run the exact same thing in ArcMap 10.7 the column names are correct in the combineRaster. Though I don't think it is the versions that make the difference in this case but what is? Something wrong with the installation? Environment variables?           ... View more