In a December 2022 "Sharing Content to ArcGIS Enterprise" Esri course, the instructor noted that when I share an ArcGIS Pro layer to Portal as a hosted feature layer, if I ensure that editing capability is not enabled, then AGE-Portal persists the layer's data as a file geodatabase instead of as an enterprise geodatabase (PostgreSQL, probably). The instructor said that the file geodatabase is the highest READ performance data structure that AGE-Portal uses for its hosted feature layers. I want to use the highest READ performance data structure.
I've only now come around to actually make use of this information, and everything changes all the time. When peeking under the AGE-Portal hood, I'm not seeing (yet, I hope not to dive into the dark box on this) that it's persisting hosted feature layers without editing enabled in anything other than its managed PostgreSQL enterprise geodatabase data store.
So, my questions:
- If, in early 2024 I use ArcGIS Pro 3.2.1 to share a layer to AGE-Portal 11.1.0 as a hosted feature layer without editing enabled, in what data structure/format is the data persisted?
- What is the highest performance data structure/format I can convince AGE-Portal to manage for me?
I have a time-enabled point feature class with ~53m rows that I need to persist in an AGE-Portal data store. Typically, only about 800 active records are queried and rendered for the full spatial envelope of the data set. Scale dependencies are set for ~1.5m active records, so they get queried and subsetted to reasonable quantities when zooming in and rendering.
There are numerous other performance, gotcha, bug, and version discrepancy red flags in my questions and examples, and I'm pretty much aware of and controlling for all of them. I'm specifically interested in understanding what controls I have over AGE-Portal's decisions on what data structures/formats to persist my data that it manages. I'll be selecting for highest possible READ performance.
Ok, well, I also want to know whether or not AGE-Portal respects and includes all of my performance configurations (field indexes, coordinate reference system, and the like) when it ingests my data into its data store. Maybe it does some of its own? I have noticed that it has not been entirely respectful of things like field order and field naming choices, so maybe it's happy to jack up or ignore my other configurations.
Cheers & thanks,
tim
edit: the only way to see typos is to glance at the post after posting it.