search

KorAP / Krill

:mag: A Corpus Data Retrieval Index using Lucene for Look-Ups
BSD 2-Clause "Simplified" License
16 stars 3 forks source link

Unexpected behaviour with Cosmas II `%-w1` #92

Open Akron opened 1 year ago

Akron commented 1 year ago

Example query: doxen %-w1 (*ortho- oder *ortho oder tho)

The %-.. operator seems to not work properly, sometimes results don't contain doxen.

I haven't checked yet, if this is a Krill or a Koral issue.

This was reported by Rainer.

Akron commented 1 year ago

The Koral looks reasonable, so I move this to Krill.

Akron commented 10 months ago

It seems this bug was also fixed with https://github.com/KorAP/Krill/commit/44491f71ef99bd47231dc55806beffcdb6b50e27

Bodmo commented 7 months ago

It seems that this bug is still there. Example: Q1: Angestelltentarife -> 18 hits, e.g. B09/FEB.02684. Q2: Angestelltentarife /s0 Geld -> only 1 hit in B09/FEB.02684: OK. Q3: Angestellentarife %s0 Geld -> 15 hits -> but should be 17 hits.

Remark: 15 hits might be correct if there were 3 hits in B09/FEB.02684 that were excluded in Q3, but this is not the case as there is only 1 "Geld" in the displayed sentence in B09/FEB.02684.

Remark: the KoralQuery of Q3 seems to be OK, as the difference between Q2 and Q3 is that the % operator is converted to exclude=true.

Akron commented 7 months ago

Ah - that's annoying! Thank you!