Workspace with more than one LRS feature dataset

[GraemeBrowning_Aurizon]

I am considering creating a new LRS for the access roads of an organization that already has an LRS for its railways.

I would like to store both of these in the same workspace (enterprise geodatabase).

I am envisaging that there will be two feature datasets in this workspace: Lrs_Rail_FD and Lrs_Road_FD and that they will share the Lrs_Centerline_Sequence, Lrs_Edit_Log and Lrs_Locks tables in that workspace.

I think that this should be fine to do but before committing to that I thought I should sound out how well multiple LRS feature datasets in the same workspace works in practice and whether there are any considerations I may be overlooking?

[NathanEasley]

We'd actually recommend against taking this approach.  While it is possible, it makes things more complicated.  Is there a reason the access roads can't be modeled as a second network is the existing LRS?

Nathan
ArcGIS Roads and Highways team

[GraemeBrowning_Aurizon]

I'm open to suggestions and recommendations Nathan

If I modelled Lrs_Rail and Lrs_Road as two LRS Line Networks within the same LRS dataset, would they each have their own routes, centerlines, redlines, calibration points and centerline sequence tables?

If I have them in separate feature datasets, I've been able to create and load these two LRS datasets which only share  Lrs_Edit_Log and Lrs_Locks tables:

From my reading of Multiple linear referencing methods—ArcGIS Pro | Documentation I'm thinking you must be suggesting that I share the centerline, calibration points and redline feature classed as well as the LRS Hierarchy between the two LRS Networks in the same LRS/Feature Dataset so that they are really just separate LRS Networks and Centerline Sequence tables.  Since the Roads and the Rail use mutually exclusive centerlines I would have thought it best to keep their LRS Networks well separated.

[NathanEasley · Accepted solution]

If the spatial references will be the same between the rail and road networks, it would make sense to model them under 1 LRS.  This would give you only 1 set of minimum schema items (centerline, calibration point, centerline sequence) to prevent duplication of items in the gdb, but each LRS Network would ensure you can still model each set of routes in their own way.