GlobalNamesArchitecture / gnresolver

MIT License
0 stars 0 forks source link

Search of non-ascii strings with ascii substitution should work #96

Open dimus opened 7 years ago

dimus commented 7 years ago

When we search "au:Aspöck" or "au:Aspock" -- both should return the same result which includes both spelling variants.

alexander-myltsev commented 7 years ago

Seems like missed DB setting.

dimus commented 7 years ago

Yes, in mysql it would be set via COLLATE=utf8_unicode_ci

dimus commented 7 years ago

Hmm, actually I think what I did is bringing all strings to "standard form" with no capitalization and substitution of non-ascii characters with ascii ones.

dimus commented 7 years ago

For example:

select * from name_strings where normalized like "%ASPOCK%"

dimus commented 7 years ago

and look at word table -- all words are normalized in mysql

alexander-myltsev commented 7 years ago

would be fixed by https://github.com/GlobalNamesArchitecture/gnresolver/issues/98