Wikidata talk:Lexicographical data/Ideas of queries
Image edit
"for all lexemes find those with изображение (P18) specified at the level of lexeme instead of level of sense" - but what if this is an image of some sign with spelling of the lexeme? --Infovarius (talk) 19:39, 4 August 2018 (UTC)
- @Infovarius: then human do not remove this image (P18) usage. Propositions of the queries are only for detection, changes belongs to human because bot do not know which sense image belongs to. KaMan (talk) 05:08, 5 August 2018 (UTC)
Uninflected words edit
I see these two rules: "for all lexemes find every lexeme without any form (by language)" and "for all verbs and nouns find lexemes with forms with empty grammatical features". This seems to imply that every verb and noun must have at least one form, and this form must have grammatical features. However, there are many languages where there are no forms other than the main form. Chinese comes to mind; Chinese does not inflect words at all. This requirement does not make sense for such languages. There is no grammatical feature that can be given to a Chinese verb, it is simply the verb.
Whether there should still be one form, but without grammatical features, I'm not sure about either. It is kind of silly if you have to create a dummy form that is identical to the lemma, for every Chinese word. —Rua (mew) 11:08, 18 September 2018 (UTC)
- Description of ideas of queries can be improved, feel free to expand it with exceptions. But it was agreed in discussions that one form per lexeme is requirement even if it is only form and identical to lemma. According to https://en.wikipedia.org/wiki/Lexeme lexeme is set of forms. KaMan (talk) 11:21, 18 September 2018 (UTC)
How to avoid timeout error? edit
Anyway to limit number of entries searched? Particularly, I need result for this query [1]. Thanks —Vis M (talk) 12:15, 13 June 2023 (UTC)
Ok, thanks. Using "LIMIT" solved my problem. Regards —Vis M (talk) 18:14, 13 June 2023 (UTC)