[Wikidata-bugs] [Maniphest] [Updated] T86278: Define which data the query service would store

2015-01-11 Thread mkroetzsch
mkroetzsch added a comment. > This is not correct, original structure can be recovered Then I misunderstood the transformation that was proposed. My impression was that a statement with three qualifier snaks: P1 V1, P1 V2, https://phabricator.wikimedia.org/P2 V3 would be stored as two statemen

[Wikidata-bugs] [Maniphest] [Updated] T86278: Define which data the query service would store

2015-01-11 Thread Smalyshev
Smalyshev added a comment. @Multichill I think there is a very good reason not to index everything. The reason is that everything has costs. If we had a system with zero index cost and infinite speed, of course, no problem. But in any real system creating indexes costs, storing indexes costs an

[Wikidata-bugs] [Maniphest] [Updated] T86278: Define which data the query service would store

2015-01-11 Thread mkroetzsch
mkroetzsch added a comment. On another note, it would be good if the view on all data that is in the system is somewhat uniform. We don't want special query syntax for badges etc. This could be avoided by viewing everything as statements, possibly using some "special" properties (and qualifiers

[Wikidata-bugs] [Maniphest] [Updated] T86278: Define which data the query service would store

2015-01-09 Thread Lydia_Pintscher
Lydia_Pintscher added a project: Wikidata. Lydia_Pintscher added a subscriber: Lydia_Pintscher. Lydia_Pintscher added a comment. Things I can think of right now - maybe more: - definitely should also import aliases. (Might already but not in the description here so making sure.) - probably also