|It is proposed that this informational page become a policy or guideline. Please see the discussion on its talk page or the project chat.|
The following is a proposed Wikidata policy or guideline. The proposal may still be in development, under discussion, or in the process of gathering consensus for adoption.
|This page in a nutshell: When merging items, be absolutely certain that the subjects are the same, then use the Merge gadget. Do not move the claims manually, otherwise the redirect will not be created. Ask if you are unsure.|
A merge is performed when two or more items exist in Wikidata on the same topic, concept, or object. A merge is a multistep process, first requiring that the collective data about the items be pooled together in one page (the recipient item), and finally resulting in the redirection of the obsolete page to the recipient item.
This guide explains how to enable automatic merges by activating the Merge gadget in your preferences and a few secondary methods.
If you have any questions about merging items, feel free to ask them on the Project Chat.
|Warning: Wikidata item IDs are designated as persistent identifiers. Therefore, merged items should be redirected. Never reuse merged items for other things.|
Check to be sureEdit
Make sure that the items are indeed about the same topic, concept, or object. There could be subtle differences requiring the existence of two or more items.
- en:tree → de:Baum (German word for tree)
- Reason: These are about the same thing (tree) and so should be merged into the (multilingual) item, tree (Q10884)
- en:tree → de:Eiche (German word for oak tree)
- Reason: These are different things (oak tree is a "subclass of" tree) and so should remain separate as two items; tree (Q10884) and oak (Q33036816), respectively
- Miller (family name) → Miller (disambiguation page)
- Reason: These are different concepts, even if some Wikipedia articles include family names on disambiguation pages. They should remain separate as two items; Miller (Q1605060) and Miller (Q304896), respectively
If you are unsure (for example, due to difficulty translating different languages), or come across items with sitelinks that appear to be inconsistent, it is best to bring up the problem at interwiki conflicts.
There are two ways to perform an automatic merge: the Merge gadget or the Special "Merge two items" page.
Please bear in mind, that while automated merges will be terminated if an interwiki conflict is found, you should still always check that the items are the same since the tools simply perform merging that you select to be done. Special:MergeItems can resolve conflicts that are merely due to redirects.
To enable this function, open Preferences, select Gadgets, check the box for Merge and hit the Save button. To use it, open either the recipient or obsolete item, go to the drop-down menu in the top bar (in the default skin, it is to the left of "Search"), select "Merge with...", and enter the ID of the other item into the merge wizard. Click on "Merge" when done.
Special:MergeLexemes can be used to trigger a merge of two lexemes if (only if) the lemma, the language and the lexical category are identical. If one of them isn't identical, it may be that this can be corrected first.
Enter the L#### in the fields specifying which is the recipient lexeme ("The ID to merge to") and which is the obsolete lexeme ("The id to merge from").
Mark as duplicateEdit
Manual merge: step by stepEdit
If there is a sitelink conflict, items can't be merged as only one link per site is possible.
Until the two pages at the site (e.g. articles at English language Wikipedia) are merged, two items have to be kept at Wikidata.
- One of the two items should be marked with Wikidata:Database reports/Identified duplicates lists all such items. and qualifier . Usually the more recent item should be marked.
- Statements, other sitelinks and labels/descriptions can be moved from the redundant item to the main item.
- Once the articles at the site are merged, the two items can be merged and deleted.
If the sitelink conflict is due to an article and a redirect pointing to the same article, the redirect can be removed and the items merged. Special:MergeItems even allows to merge them directly.
Finding items to mergeEdit
Now that you know how to merge, you can make use of your new skill at various places on Wikidata where potential duplicates are found/listed. These places include:
- Wikidata:The Game
- Wikimedia tool which has gamified the merging process, offering candidate pages to potentially be merged by the user
- User:Ivan A. Krestinin/To merge
- lists of merge candidates based on similar page title strings
- lists of merge candidates based on sitelink names being identical
- Commonsmerge and Uniquemerge
- lists of merge candidates based on values of a property being same in two items
- Bene's Lonely Items tool
- finds items containing only one sitelink to a Wikimedia site (which therefore may need to be merged with other items linked to different Wikimedia language versions of the same page)
- Not in the other language
- lists items in a given category that are linked to articles in Wikipedia A but not in Wikipedia B
- Constraint violations
- summary of constraint violations for properties; in particular, "Unique value" violations may indicate a duplication
- Wikidata:True duplicates
- list of items having at least one sitelink which is due to a bug used in another item
- Wikidata:Database reports/identical birth and death dates
- based on date of birth (P569) and date of death (P570)
- Wikidata:Database reports/Identified duplicates
- items with Wikimedia duplicated page (Q17362920) waiting for Wikipedia to merge
- Wikidata:VIAF/cluster/linking to multiple Wikidata items
- items which, according to VIAF (see VIAF ID (P214)), are related to the same person - VIAF may sometimes be wrong
Other pages and projects in the Category:Merge candidates
Feel free to add any more useful pages or delete ones that are not actively updated!
Canceling a wrong merge could be done in a few steps:
- Go to the history page of the item onto which the merge has been done and restore the revision just before the merge.
- Go to the history page of the merged item and restore the revision just before the merge.
The order is important (especially if the item has sitelinks) because articles can be linked to at most one item. So they should first be deleted on the item which got all the linked before they could be restored on the other one.
Some bots (KrBot for example) are known to replace the redirected items in statements with their target redirects a certain period of time (24 hours for KrBot) after the merge. The next step is to ping bot owners (like Ivan A. Krestinin) so they can run the code to cancel the if that period of time passed and some bot did the substitution, so that the statements can be restored with the right item. KrBot creates edit groups when fixing redirects, which may be reverted en masse using the EditGroups tool.
On a closely related topic, the help page Help:Split an item will help you if the item has not been merged but if statements on two topics are mixed-up on a unique item for some reason. It has guidelines on how to divide the statements on two items without re-entering the information by hand.
For related Help pages, see:
- Help:Sitelinks, which explains what sitelinks are and what rules they follow
- Wikidata:Interwiki conflicts, for reporting interwiki conflicts
- Wikidata:Requests for deletions, for nominating items for deletion
- Wikidata:Tools, freely available tools for working with Wikidata
For additional information and guidance, see:
- Project chat, for discussing all and any aspects of Wikidata
- Wikidata:Glossary, the glossary of terms used in this and other Help pages
- Help:FAQ, frequently asked questions asked and answered by the Wikidata community
- Help:Contents, the Help portal featuring all the documentation available for Wikidata