Add Join — Pseudo-ObjectID (row number) as unique ID field for 1:M join

Bud

Regarding BUG-000154798: The 'Select By Attributes' tool does not select only records with specified field values added through a one-to-many join but instead selects all records with duplicate ObjectIDs. (Status:  Known Limit)

• Select By Attributes on joined data — For rows that are 1:M, all rows in the join table get selected, despite selection criteria
• Is ArcGIS Pro the right tool for tabular/join-based geodatabase analysis?

Could the various issues with 1:M (one-to-many) joins in Pro be solved by having a pseudo-ObjectID (row number) field as a unique ID?

For example, Oracle's ROWNUM pseudo-column works fine for query layers as the unique ID. Could a similar row-number-based mechanism work for Add Join in the attribute table, too?

select
    cast(rownum as int) as pseudo_objectid,
    lc.*,
    pf.*
from
    lc_events lc --one
left join
    lc_events_project_finances pf --many
    on lc.event_id = pf.event_id

Edit: I think this approach would work for M:M (many-to-many) joins too.

Related: Disallow 1:M join layers as GP inputs instead of producing unexpected results

JonathanNeal

@Bud Yes, internally selections work off of objectId's, stored in the layer.  When there is a 1:many join, selections will select duplicates.  For the Add Join tool, we would require additional development by the Geodatabase team. 
The team does have an issue for this:
 Title: "Relquerytable to support a unique generated field"
Note1: this is a tough task when implementing it across all data sources, so it will require lots of considerations.
Note2: this will also make analysis tools work correctly on joined layers.