ArcGIS Pro 2.6.8; Oracle 18c 10.7.1 EGDB (edit: SDE.ST_Geometry):
I have feature classes that have rows where the shape is null. If I create a query layer on one of those FCs, the query layer will automatically hide the rows that have null shapes from the attribute table.
Video:
It's confusing when the query layer attribute table has fewer rows than the underlying datasource. The user isn't warned that the rows get automatically hidden. We also have specific cases where we want to see null shape rows in the attribute table so that we can investigate them. Like this:
https://community.esri.com/t5/arcgis-pro-questions/export-fc-while-maintaining-objectid-values/m-p/1...
Null shapes are perfectly valid. We can create them using the attribute table, field calculator, python, and SQL. For example, we have a workflow where we load data into a FC from a non-spatial source (like a spreadsheet) and then populate the shapes after the fact. If null shapes are valid, then I don't think query layers should hide them from the attribute table.
To be honest, I've always had trust issues with query layers. They often silently exclude data, which makes analysis problematic. So, I end up avoiding ArcGIS Pro for queries and using other platforms that don't silently exclude rows. Improving things like this would go a long way in terms of making query layers more reliable.
Note: I've only talked about null shapes here. But this idea also applies to features where the shape is not null, but the geometry is nil or null. See states #2 and #3 here: Select features that have blank shapes. Such features should also be included in the attribute table.
