Hi. I was wondering how I can find texts with asterisks(*) or question marks(?) in it.
Using the following operator doesn’t seem to work as they act as wildcards.
{
path: ["text"],
operator: Like,
valueText:"*"
}
And is there any reference of special characters that require such handling other than asterisks and question marks?
Thanks in advance.
Hi! This is an interesting question
I will ask internally.
A side note is that, if your tokenization is word
(de default one), it should keep only alpha numeric values:
jphwang
February 26, 2024, 3:35pm
3
Hi @Jinhyun_Bang - sorry about the late reply. This one slipped through. Currently, the asterisk (*
) character cannot be searched for with Like
unfortunately.
This would require implementation of a custom tokenizer, which we are evaluating for non-English language use for example (e.g. this thread Chinese tokenization support · Issue #3102 · weaviate/weaviate · GitHub ).
Sorry about that. Hopefully we’ll have that available in the future.
1 Like