Decide where the work belongs
Identify the relationship consequence
Compare intended parent-dependent pairs with final links in controlled examples. Record rejected dependents, accepted but unlinked records, and later resolution separately. Keep downstream workflow effects tied to demonstrated relationships rather than assuming every unresolved reference changes the same behavior or that eventual parent presence guarantees automatic repair.
Scope the order-sensitive population
Group affected imports by source layout and whether parents arrive in the same file, another file, or an existing dataset. Compare a known supported order as a control. Estimate reach from those patterns rather than all relational imports, and separate identity mismatches that remain wrong under either processing order.
Assess recovery and guidance
Try the documented order or deferred-resolution flow with the controlled pair. Determine whether users can identify and retry unresolved links without repeating accepted work. Include manual sorting and cross-file coordination effort, especially when the source export cannot naturally guarantee the processing sequence the importer currently requires.
What to carry forward
Record priority with confirmed relationship loss, source-pattern reach, silent-acceptance risk, and recovery cost. Assign a trigger for unlinked records presented as complete. Keep ordering guidance separate from a behavior guarantee; a recommended file sequence is only a reliable workaround when the importer consistently honors the relevant dependency boundary.
Technical background: PostgreSQL documentation.
Keep the decision with the work.
Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.
Create an account See the product workflow