Page MenuHomePhabricator

Inconsistent query results for recently created items on query service
Closed, DuplicatePublic

Description

I'm importing paintings for the National Gallery of Canada (see https://www.wikidata.org/wiki/Wikidata:WikiProject_sum_of_all_paintings/Collection/National_Gallery_of_Canada ). In the first run I imported a bit over 4000 paintings, but I skipped a couple. So I'm doing a second run to fill the holes. The bot makes a lookup table based on the collection and inventory number

SELECT ?item ?inv  WHERE {
  ?item p:P195 ?collectionstatement .
  ?collectionstatement ps:P195 wd:Q1068063 .
  ?item wdt:P31 wd:Q3305213 .
  ?item wdt:P217 ?inv . 
  } LIMIT 25000

Based on the lookup table the bot will expand existing items instead of creating new ones.

I noticed duplicate items getting created. See https://www.wikidata.org/wiki/Q59492235 and https://www.wikidata.org/wiki/Q59551206 . When I manually run the query I sometimes get 4011 results and sometimes 4002 (just change the limit to something like 1234567 to trigger a fresh run). So the data seems to be inconsistent between member servers.

Event Timeline

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptDec 8 2018, 11:13 AM
Multichill triaged this task as High priority.Dec 8 2018, 11:13 AM