I'm wondering if anyone else out there has experienced a similar issue with 10.3.1 or if someone from Esri can duplicate my problem...
I recently received a new machine and installed 10.3.1 on it and I seem to have an issue with feature classes not having the same SRID as the dataset they are contained within. If I create a similar dataset with 10.3 and do the same steps, I don't have this problem.
Below are some details of the issue:
1) Dataset Working103 = dataset created with ArcGIS 10.3. SRID = 103045 (NAD 83 Iowa State Plane North 2011)
2) Dataset Working1031 = dataset created with ArcGIS 10.3.1. SRID = 103045 (NAD 83 Iowa State Plane North 2011)
Next, I created a simple point feature class within each dataset, and inserted 3 features within each feature class. The feature class within the dataset created with 10.3 has the same SRID as the dataset, 103045. The feature class within the dataset created with 10.3.1 has a SRID of 6463, which I can't seem to find browsing through Esri's list of coordinate systems, so I have no idea what it is.
Feature class within dataset created with 10.3 has the correct SRID
Feature class within dataset created with 10.3.1 doesn't have the same SRID as the dataset
If I do a spatial intersect query within Sql Server on the 2 feature classes using a polygon feature class from our production gdb, where all feature classes have a SRID of 103045, the feature class with the correct SRID returns results, while the other does not.
Spatial intersect query works with dataset created in 10.3 bc SRIDs match
Spatial intersect query does NOT work with dataset created in 10.3.1 bc SRIDs do NOT match
Has anyone else experienced this issue...? Is this a known bug...? Is this specific to our use of SRID 103045...?
It's due to the fact that the CRS had an Esri number at 10.3.0, 103045, but was "code changed" at 10.3.1 to update the well-known ID to the EPSG one, 6463. They're just different numbers pointing at the same definition. Using the older number in a query that passes through the Esri projection engine, the older number is converted to the new one before use.
I'm not sure what's occurring on the DBMS-side.