I have some unemployment data, and in its raw form it is a table with unique census blocks as rows, and a series of columns for different date periods and the corresponding unemployment rate in those time periods.
I munged the data to flatten out the date fields into a single field and have multiple records for each census block, in line with the 'temporal data in separate tables' help link and this tech session.
As the census blocks never change, but the attributes change for each date, I figure it makes sense to store the data in a 1:M relationship, so that when I am working with a service in a web app, I won't have to keep requesting the same geoms (census blocks) for each time interval.
Now I have this data in a standalone table:
- I then used the Add Join tool against the corresponding census block featureclass (1:M)
- Enabled time on the Date field
- Can playback through the data in Pro
- Joins do not appear to flow through when publishing to ArcGIS Online
- However, publishing as a map image layer to ArcGIS Enterprise seems to be ok, but was hoping to just rely on AGO
- Adding the feature service to the Map Viewer does not let me symbolise on the field from the joined table
- But via the Map Image layer, I can symbolise on the unemployment_rate field from the joined table:
However, I don't want to include ArcGIS Enterprise in the equation and also want to work with featurelayer and not MapImageLayer so I can have more interactivity with the final app. But I do not want to create a flattened layer with duplicate geoms for every time interval, as this will mean the client app has to keep loading unnecessary geoms for each time interval (i.e. Census blocks don't change for each time interval, but the attribute values do).
So the crux of my issue is how to serve up 1:M temporal data to ArcGIS Online?