Select to view content in your preferred language

Is there any functional difference between a feature class and a SQL table with a geometry column?

1525
3
Jump to solution
07-27-2021 12:17 PM
JVig
by
Occasional Contributor

Hello,

I am wondering if there is any functional difference between a feature class created using any of the arcpy feature class related methods versus building a table in SQL Server, enabling it with geometry, and ensuring a unique primary key. While I understand that tables created in the back end would not be registered to the geodatabase, after registration is there any functional difference?

 

Thank you!

0 Kudos
1 Solution

Accepted Solutions
JVig
by
Occasional Contributor

For any who are interested, my testing has shown that there is little difference in the behavior of the tables once they are registered to the database. You can effectively create a data ingestion workflow to bypass the use of, say, a GP Server and arcpy, to load datasets from source. If the data is hosted as a service, the service will update automatically as the underlying tables update.

View solution in original post

0 Kudos
3 Replies
JVig
by
Occasional Contributor

For any who are interested, my testing has shown that there is little difference in the behavior of the tables once they are registered to the database. You can effectively create a data ingestion workflow to bypass the use of, say, a GP Server and arcpy, to load datasets from source. If the data is hosted as a service, the service will update automatically as the underlying tables update.

0 Kudos
BrentHoskisson
Frequent Contributor

You kind of need to define what you mean by Feature Class.  If you are meaning SDE Binary vs. SQL Spatial, it turns out that there is some indexing differences that makes SDE Binary ultimately run a tad faster than SQL Spatial.  You only START noticing if you have layers with over 500000 or so rows.  

Both have full functionality in Arc products as far as I have determined.

JVig
by
Occasional Contributor

I  guess this response really nails the true question. What is a feature class? Is it just a table with a geometry/geography field and a unique id?

We source data from our data lake and feed it through the pipeline back to an RDS. I was noting that a gp service had been created to do the ingestion portion of this and it was occurring on the same server customers were hitting to consume services causing periodic slowdowns. However, we have existing infrastructure to lift from source to db for non-spatial data and I wanted to leverage this to bypass a gp ingestion process and then serve the data as a service. We have managed to construct this pipeline effectively but I was wondering at the core, what is a feature class? Does the construction of a feature class in purely sql terms (without using the proprietary arcpy feature class creation methods) affect the service? I have not noticed anything out of the ordinary but I have verified we are using SDE binary. I am thinking there isn't much of a difference at all honestly but this is all sort of new for me 😄

0 Kudos