After importing an item from wikidatawiki to testwikidatawiki it breaks[1]. The item is now editable like a normal page.
[1] https://test.wikidata.org/w/index.php?title=Q1000000&action=edit
Version: unspecified
Severity: normal
After importing an item from wikidatawiki to testwikidatawiki it breaks[1]. The item is now editable like a normal page.
[1] https://test.wikidata.org/w/index.php?title=Q1000000&action=edit
Version: unspecified
Severity: normal
Status | Subtype | Assigned | Task | ||
---|---|---|---|---|---|
Declined | None | T53412 Import of items to other MediaWiki installations with Wikibase not possible | |||
Resolved | None | T49070 It is possible to import standard wikitext pages into the main (item) namespace |
Interesting, Personally I import from xml to my local test install rather a lot and that seems to work fine, but not when importing directly from wikibase!
the import will appear as if it is an entity to start with but the only way to edit is using the standard edit tab. One edit using this and the page breaks simply displaying the json!
Same problem with XML upload[1] (first example was via transwiki).
[1] https://test.wikidata.org/w/index.php?title=Q600000&action=history
I've been thinking about this and I don't see a way to do a meaningful import that does what the user expects.
When you import items you will mess up all of the references to other items which don't have the same IDs on another system. In most cases you will even get links to items that are about a completely different topic and just happen to have the same ID as the item you really wanted on the old system.
@Lydia_Pintscher , that is not dissimilar to importing wikitext pages, especially the process of importing templates, where a strict order is needed, but is not provided . So admins just import them one by one until they all work.
However I am pretty sure this bug was fixed at some point, as I know I did this import (using special:import) with a local Wikibase in 2014, and it definitely did work.
Anyway, I see we have T111787 and T90148 about this.
IMO it is critical that Wikimedia not just export dumps, but ensure the db can be easily rebuilt from those dumps. (actually this is a GPL legal requirement, which is often overlooked, and to clarify, last time I tried loading a wikidata dump it did work, so all good here) And the "one page at a time" via special pages strategy for exporting/importing is a real necessity at times.