Decide where the work belongs
Identify the wrong assignments
Compare intended and saved destinations for controlled records from each source file. Record swapped fields, missing values, and defaulted columns separately. Determine whether downstream validation can detect the mismatch; similar text fields can accept each other’s values and leave the overall import looking successful despite incorrect meaning.
Scope schema variation in the workflow
Group affected files by producer, version, reordered columns, and optional fields. Compare batches with consistent schemas as controls. Estimate reach from actual mixed-layout batches rather than all multi-file imports, and retain uncertainty when accepted records no longer carry a source reference that would support reliable attribution.
Assess review and reconciliation
Check whether users can inspect each distinct schema before confirmation and whether the chosen alignment rule is clear. Estimate the effort to trace already accepted records back to their source files. A workaround that requires manually reordering every export may be fragile when the producer changes optional columns between runs.
What to carry forward
Record priority with confirmed assignment errors, mixed-schema reach, detection limits, and recovery effort. Assign a trigger for silent acceptance from another layout. Keep the decision focused on a predictable alignment contract rather than simply increasing the number of rows shown from the first file in the preview.
Technical background: DuckDB 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