Re: [basex-talk] Full text search with non-Latin scripts

Hi Andreas,

Sorry for the late reply.

Fuzzy/approximate search is currently limited to the Latin character set.

Best regards, Christian

Andreas Hartmann andreas.hartmann@gmx.org schrieb am Sa., 17. Sept. 2022, 12:33:

Hi,

I have problems with full text search in 10.1, when non-Latin script is involved. Expressions such as

"βάναυσος" contains text "67644567" using fuzzy "67644567" contains text "βάναυσος" using fuzzy "humbug" contains text "βάναυσος" using fuzzy

all yield "true". It's the same with Hebrew strings. It does not matter which Greek/Hebrew word is used, as long as the word-length does not differ too much from the compared string. Is this really the expected behavior or could it be a bug?

Best greetings Andreas