Multilanguage search without text search

Working with Data
1

Hi everyone, I’m currently developing an app with mongo as db. In the app I have certain documents that have names in different languages and users can search them with a textbox with autocompletion.
Once the documents are found, the user can either use them to build up another document or “import” them in their private collection. Here the collection is not ment as a mongo collection, in fact what tells me that a document was imported is the field “userId”, which when import edis set to the user id who imported it or set to a generic userid which represent the “public user”. My queries are the following:

{"userId" : <userId>} -> give all imported documents if userId = userId or all public document if userId = publicUserId
{"userId" : <userId>, "name" : <keywords>} -> search by name using keywords (imported = userId, public = publicUserId)
{"userId" : <userId>, "tags" : <list of tags>, "name" : <keywords>} -> same as before, but some tags can be added

Now, the structure of my document is the following:

{
     _id,
    uniqueId,
    name : {
        it: <name it>,
        en : <name en>,
        ...
    },
    nameSearch: {
        it: [
            <token 1>,
            <token 2>,
            ....
        ],
        en: [
            <token 1>,
            <token 2>,
            ....
        ]
    }
}

This works fine as I can match the full name or a part of it with the following clause:

$and : [ 
         {
             "nameSearch.en": {
                 $regex: "^<keyword1>"
             }
         },
         {
             "nameSearch.en": {
                 $regex: "^<keyword2>"
             }
         }
      ]

And I have the following indexes to support the query:

{"userId", "nameSearch.it"}
{"userId", "nameSearch.en"}

My question here is, wouldn it be “better” to put the tokens of all languages in a single list? For example, lets say that the name is “Documento importante” in it and “Important document” in en, instead of having a list of token for each language I will put all the tokens in the same list:

nameSearch : ["documento", "importante", "document", "important"]

That way the user can automaitcally search by all the languages without having to instruct the back end to do so.
I know I can use the full text search, but I was doing some trials on it and I found that the index is bigger and also slower.

Thank you

Green