Keep near-label review pairs from becoming one merged asset

Preserve the archive’s working references

A fictional studio is preparing an archive handoff for a recurring campaign. A cleanup suggestion groups three records under one heading and proposes retaining only the most recent preview. The archive editor has not yet examined the underlying artwork. The producer’s job is to return a useful review sheet, not to decide that a cleaner-looking list is a more accurate collection.

The records may lead to different approved editions, working files or presentation material. None of those roles can be established from a convenient preview alone. Begin with the supplied references and keep them visible on the review sheet. Give the reviewer a place to record what each asset contains, the campaign use it served and the evidence available for its status. Leave unknown information blank rather than filling it from a filename guess.

This task is deliberately upstream of a final archive decision. It does not replace the editor’s comparison of content, approval history or rights. It should make that comparison easier to perform without silently removing the very record that might explain a difference. A good cleanup aid directs attention; it does not manufacture certainty by hiding alternatives.

Inspect the entire candidate construction

Use Levenshtein distance: the minimum number of unit-cost character insertions, deletions or substitutions. Exact record IDs R1, R2, R3 have compared labels AB12, AB13, AC13 respectively. The three pair distances are R1–R2: 1; R2–R3: 1; R1–R3: 2. Draw a candidate edge only when distance ≤ 1. R1–R2–R3 is connected, but its endpoints fail that rule. Reachability therefore does not establish an all-pairs match or authorize merging. Keep every ID and label unchanged.

The displayed results were calculated locally in Python and independently checked by enumerating possible edits. No PostgreSQL database or archive service was run. The documentation grounds the distance convention; the exercise does not claim that a production system uses that convention or that its output has been inspected.

Place the complete comparison beside the original cleanup suggestion. Do not replace the suggestion with a new unexplained verdict. The reviewer should see both what made the suggestion plausible and what it leaves unestablished. This makes the correction constructive: the aid can remain useful without being granted authority it was never designed to have.

Give the reviewer the underlying asset references before presenting a polished consolidated preview. The order matters editorially because a single attractive image can make the proposed cleanup feel already decided. Start with the evidence that still needs examination, and show the original proposed grouping only as the question being investigated.

Give each proposed relationship its own disposition

Add separate evidence notes for the relationships being reviewed. Useful evidence might include an approved manifest, an explicit revision history or a content comparison by the responsible editor. Describe what the evidence actually establishes. A shared campaign name does not establish that two files serve the same purpose, and the newest file date does not identify the approved edition by itself.

The studio can now distinguish several ordinary outcomes without taking an external action. A record might contain an obsolete draft that still needs historical retention. Another might be the current approved deliverable. A third might be an alternate composition made for a different placement. These are possible review findings, not facts inferred about the three fictional labels. The exercise supplies no artwork from which to select one of those findings.

If the reviewer confirms a naming error, record the proposed correction separately from any decision about retention. Fixing a display label and discarding a record are different actions with different consequences. Likewise, an approved relationship between editions should not erase which exact file was delivered for an earlier campaign. Keep the explanatory history available to the person who must answer a later client question.

Where evidence is missing, the disposition remains unresolved. That is a usable result: the next reviewer knows what to inspect and which question remains open. Do not invent a relationship just to make every row look finished. The archive handoff can report incomplete review without presenting the collection as defective or implying that a hidden automatic cleanup has occurred.

Choose a review aid rather than a reduced collection

The selected deliverable is the review sheet, not the proposed single replacement record. It includes the untouched references, the complete supplied comparison, evidence notes and an explicit unresolved status where necessary. The archive editor receives a manageable task while retaining access to all of the material that may matter.

The cost of this choice is review work. A producer cannot advertise the proposed cleanup as completed merely because the sheet is ready. Explain that distinction in the handoff: the candidate review has been prepared, while any consequential archive action still requires its own evidence and authorization. Do not describe this as a measured productivity gain or promise that the review will discover a specific number of redundant assets.

For future campaign handoffs, ask that the approved manifest identify the actual deliverables and their purposes. That improves the context available to the next editor without requiring a speculative rewrite of old history. The present exercise ends with all original records retained and the cleanup suggestion made inspectable. Its value is a sharper review question, not a smaller record count.

Make the reasoning concrete.

A campaign cleanup suggestion offers one recent preview as the replacement for several unexamined records.

Choose the review sheet and keep unresolved findings visible. No real asset is renamed, deleted or consolidated.

Take the questions into your next brief.

Use these prompts in your own notes. Nothing is submitted from this page.

1. References

   Can the reviewer still locate every original record?

2. Comparison

   Is the complete supplied comparison visible beside the suggestion?

3. Evidence

   What establishes the actual role of each asset?

4. Disposition

   Which questions remain unresolved rather than assumed?

5. Handoff

   Does the status distinguish prepared review from completed cleanup?

Watch for the wrong conclusion.

• Treating the newest preview as the approved deliverable without evidence.
• Removing an unresolved record to make the handoff look finished.
• Combining a proposed label correction with an unapproved retention decision.

Sources & scope

The archive scenario is invented. Local computations only; no database, live archive or record consolidation was operated.

• F.16. fuzzystrmatch — determine string similarities and distance →Primary PostgreSQL documentation · Supports: Distance convention.