Searching based on Emoji


(Gautam S Thakur) #1

I am interested in writing a set of queries that will search for emoji in the text (data already in the ES). What should be appropriate string to use to search in ElasticSearch? For example, I want to search for 'happy face' (:grinning:). What is the appropriate query string would be -- below (replace "?????" with query string). Thank you.

GET /myindex/data_stream/_search
{
"query": {
"match": {
"text": "?????"
}
}
}


(Ivan Brusic) #2

I would assume that emojis are saved with their unicode values:
http://unicode.org/emoji/charts/full-emoji-list.html

You would need to verify the contents by looking at your data. Make sure
that the unicode string was not tokenized into smaller tokens.


(Gautam S Thakur) #3

I don't think unicode string is tokenized. How to confirm that anyways? Thank you very much. I tried searching with unicode. doesn't seems to work


(Gautam S Thakur) #4

So, I tried using corresponding unicode characters. ES returns no results. Despite the fact that I can see it has relevant data. Attached screen shots shows it (when "smile" is used instead of unicode. Just because it is part of the text) -


Check out sense result screen. I can see emoticons. But when tried with their corresponding unicode (\u263a and a couple of other also that I tried), ES didn't return any result. What would be the right query parameters, so I can search emoticon/emojis? Thanks.


#5

Hi Gautam,

Any luck? I am also in keen on the same!


(David Pilato) #6

@flamboyantn better to open a new discussion.

Anyway, have a look at what @Damien_Alexandre wrote: