Find and Resolve Duplicates
Same as Comparing Objects, which this page says it reuses: an exhaustive
search of both starhive-development and starhive-client found no duplicate-finding/merge
feature (Duplicate, Dedup, MergeGroup, FindSimilar, and similar terms, plus UI copy like
"Set as original") in either repo. It may live elsewhere or no longer exist — treat everything
below as unconfirmed until checked against a live instance.
Find Duplicates helps you clean up objects that are really the same record — the same company entered twice, the same person under two spellings. You define what makes two objects the same, Starhive finds the matches, and you decide what to keep, group by group.
You work on one type at a time, reached from that type. Searching changes nothing by itself; you make every change yourself. Merging requires permission to manage objects in the space — without it, you can still search and read through what was found.
Step 1 — Define a duplicate
Only the first choice below is required; the other two sharpen the search.
| Choice | What it means |
|---|---|
| Attributes | Two objects are duplicates when every attribute you pick here has the same value — pick the ones that identify a record (an email, a registration number, a name). You can pick inherited attributes as well as the type's own; images and files aren't offered. |
| Identify similar objects | Also groups text values that merely read alike, like "Acme Inc" and "Acme Corporation" — useful for hand-typed names. Applies to text attributes only; dates, numbers, and references still have to match exactly. |
| Narrow the scope | Limits the search to part of your data (e.g. only active records). Leave empty to search every object of the type. |
Capitalization and surrounding spaces are ignored either way — "acme inc" and "Acme Inc " count as the same.
Pick at least one attribute, then Find duplicates. If nothing matches, you stay on this screen and can widen your choices.
Step 2 — Review the groups
Each row is a set of objects that look like the same record.
| Column | What it shows |
|---|---|
| Group name | The matching values, so you can tell at a glance what the row is about. |
| No. of objects | How many objects remain in the group — decreases as you delete duplicates. |
| Status | Not resolved until every duplicate in the group has been reviewed, then Resolved. |
| Action | Resolve opens the group; once finished it reads Review, and you can reopen it. |
Progress is remembered while Find Duplicates stays open; closing it with unresolved groups asks for confirmation first.
Merging a group
Resolve opens two objects side by side — the one you're keeping on the left, one duplicate on the right — one duplicate at a time, with a progress bar at the bottom.
The kept object is the original; Starhive starts with the oldest object in the group, but you can click Set as original on a different one if it's the better record (the previous original then becomes a duplicate to review).
- Compare the two sides — matching attributes sit opposite each other.
- Copy anything worth keeping from the duplicate onto the original, one attribute at a time, using the button between the two cards.
- Delete the duplicate once you've taken what you need, or leave it and move on — both count as reviewed.
- Repeat until the group has no duplicates left; it's then marked resolved.
A copy applies to the original immediately — there's no separate save. Set as original and Delete are unavailable without permission to manage objects.
Good to know
- Results are a snapshot from when you searched — objects added or changed afterward aren't included, so search again for a fresh picture.
- Starting a new search clears progress from the previous one.
- Start narrow (one or two strongly identifying attributes, exact matching) and only loosen the search afterward — a wide search groups records that are merely alike, and every merge deletes something.