This message was deleted.
# community-help
s
This message was deleted.
f
Have you set the locale parameter on the field?
v
yes
Copy code
.locale(MARATHI_LANGUAGE_LOCALE)
f
What's the value of the
MARATHI_LANGUAGE_LOCALE
?
v
"mr" its for Marathi. Its an Indic Language
f
Hm, maybe it's an issue with the ICU tokenizer. @Kishore Nallan may have come across something similar in the past
v
i have changed my approach to use the 1st token "*संस्कृती*" as a single query input, retrieve the matched documents and then create a new collection with those retrieved documents to query using the 2nd token "*शंकर*" and that is giving me consistent desired result throughout different test cases. is there a better approach or I should stick with this? 😅
f
I would guess your best bet is to do what fixed the problem, sadly we're depending on the tokenizer based on the locale
v
Thanks. Appreciate the help 👍
🙌 1