How-to
Reviewing Duplicate Records
Reviewing the possible duplicate person records the system finds every day, merging records of the same person and taking different people out of the queue.
The system scans person records every day and lists pairs that may belong to the same person on the Duplicates tab of the Catalogue → People & institutions screen. For each pair you merge the records or mark them as different people. The tab is open only to people with the merge permission.
Building the list
- Pairs of people with similar names go on the list. The percentage on the row shows how similar the two names are.
- Pairs with a similarity below 55 per cent do not go on the list.
- Two records with the same ORCID iD count as 100 per cent similar.
- If two people share an institution, 10 points are added to their similarity.
- People with different ORCID iDs never match.
- Deleted and merged people are not scanned.
- A pair that was merged or marked as different people does not come back to the list. The scan counts only new pairs, so a new pair is found however many old pairs fill the queue.
- When an import, a DergiPark harvest or a release brings a new author, the system does not guess a match with a person of the same name. If the author's ORCID iD is not in the catalogue, or several people have that name, a new person record is created and enters the list with its namesakes at 100 percent similarity. You decide whether the records belong to the same person.
The list is sorted from the highest similarity down.
Reviewing the pairs
- On the Catalogue → People & institutions screen, open the Duplicates tab.
- Compare the names, ORCID iDs and institutions of the two people on the row.
- If you are not sure, press a person's name to open their page.
- If the two records belong to the same person, press one of the two merge buttons on the row.
- In the window that opens, complete the merge as in Merging Person Records.
- If the records belong to different people, press Different people.
A merge button shows the two names with an arrow. The record on the left of the arrow is removed, and the record on the right is kept. A pair you mark with Different people leaves the queue, and the people's records do not change.
Running the scan at once
To scan newly added people without waiting for the next day, press Scan again at the top of the tab. The notification says how many new possible duplicates were added to the queue.
A person's pending matches also appear on the Possible duplicates card of their own page. The Merge button on the card opens the window with the person whose page you are on as the source and the matching person as the target. If needed, turn the direction round with Swap direction.