A movie search can return two distinct works with the same title and several disc editions of each. If a catalog treats the title string as its identity key, a harmless-looking deduplication rule can combine the wrong records and send a shopper to the wrong product.
This is a proposed data and UX design for that problem, not a report of a deployed system. The examples below use invented records so that the behavior can be examined without making claims about real inventory.
Separate the work from its names and sellable editions
Give each film work a stable internal identifier. Attach titles as names associated with that work, with fields for language or territory when known. A localized title or spelling variant should help retrieval without creating an extra film automatically.
Keep a physical edition in a separate record. It points to the work and carries its own product identifier, format, and verified edition details. If cuts need separate records, model those between the work and the physical edition rather than hiding them in the title field.
For a small design fixture, imagine work F101 called The Lantern, released in 1984 and directed by Mara Vale. Work F202 has the same display title, a 2016 release year, and director Owen Reed. Edition E301 is a DVD of F101; E302 is a Blu-ray of F101. Those two editions belong together under one work. F202 remains a different work.
The names, years, and people in that fixture are fictional. Its purpose is to make the identity boundary visible.
Normalize for matching, not for merging
Case folding, whitespace cleanup, and punctuation normalization can make a query easier to match. They do not establish that two records describe the same film. Store the original display text and use normalized forms only for retrieval.
Even title plus year should be treated as candidate-matching evidence rather than a guaranteed unique key. Release years can reflect different events or territories, and incomplete imports can omit the distinguishing details. Conflicting records need review, not an automatic overwrite of whichever value arrived first.
A remake relationship should be an explicit, sourced association between work identifiers. Never generate that relationship merely because two titles match.
Put disambiguation beside the title
For a title-only query such as The Lantern, show separate work results. Each result should put the film year and director next to the title, with a small edition summary below. Label a missing director or year as unavailable instead of silently borrowing it from another result.
The link's accessible name should distinguish the works too. A label such as "The Lantern, 1984, directed by Mara Vale" provides more context than two links both named "View movie." Cover artwork can supplement this information, but it should not be the only way to choose.
If the query includes a year or director, rank the matching work more prominently while retaining a clear path to other results. Explain a suggested alternative rather than silently replacing an unmatched year with a nearby one.
Preserve selection through the purchase flow
Once a visitor chooses F101, opening a DVD or Blu-ray option should keep F101 selected. The cart should carry the exact edition identifier and display enough film information for the visitor to recognize the choice.
A rerelease date belongs to the edition and should be labeled accordingly. Substituting it for the film year would recreate the original ambiguity at checkout.
A public catalog such as DVDWholesaleShop's US shop supplies the browsing context for this design discussion. The proposed identifiers and behaviors above are illustrative and do not describe that site's current implementation.
Test the boundaries that are easy to lose
Use a fixture containing two same-title works, multiple editions of one work, one alternate title, and a record with an unknown year. A title-only query should retain both works. An alternate-title query should find its existing work. Selecting a format should not change the work identifier.
Also check that a missing year stays missing, that a disc rerelease date never becomes a film year, and that two matching title strings do not create a remake relationship. These are useful acceptance checks because each catches a distinct failure that a successful search response alone would miss.
The central design decision is small: a title is a name, not an identity. Keeping that distinction intact makes search, edition selection, and the cart easier to understand.
Top comments (0)