MongoDB Text Index Error: language override not supported

Question

I am using version 2.6.1. I am trying to create a text index and I'm getting error:

{
    "connectionId" : 4932,
    "err" : "language override unsupported: en-US",
    "code" : 17262,
    "n" : 0,
    "ok" : 1
}

The documents being searched have a "language" field that has a value of "en-US", but it is not used to override the language in a text search. I tried to create the text index to specify a field that doesn't exist ("lang"); however, I get the same error. I was able to create the index just fine on version 2.6.0. Is there a way to create the text index and ignore the language_override field?

Here is the working command I used on 2.6.0 (doesn't work on 2.6.1):

db.collection.ensureIndex({ title: "text" }, { name: "TextIndex" })

Here is the command I tried on 2.6.1 to specify another language_override field that does not exist:

db.collection.ensureIndex({ title: "text" }, { name: "TextIndex" }, { language_override: "lang" })

Thanks in advance!

Answer 1

Solution:

Set the default_language and language_override to the same literal value (in your case "en").

How I got here...

I hit the same problem, also on Mongo 2.6.1.

In my case I created the index with a language_override pointing to a language field where there were already documents with unsupported values (e.g. 'ar' - Arabic).

Here's how I was creating the index:

db.users.ensureIndex({ 
  "displayName": "text", 
  "about": "text", 
  "email": "text" 
}, { 
  "name": "users_full_text", 
  "default_language": "en",
  "language_override": "language"
});

I was hoping it would fall back to the default_language when the language_override value is unsupported, but apparently not. Here's what Mongo says:

{
  "createdCollectionAutomatically" : false,
  "numIndexesBefore" : 3,
  "ok" : 0,
  "errmsg" : "language override unsupported: ar",
  "code" : 17262
}

OK, fine, so the index wasn't created but I should be able to create it without the language_override, right? Wrong - mongo gives me the same error even though I no longer have the language_override specified.

The failed attempt at creating the index seems to have left behind some broken version of the index that does not show up anywhere so I can't drop it (it doesn't appear in db.users.getIndexes() and dropping it by name doesn't work).

In the end I managed to fix the index by setting the language_override to the literal value 'en', like this:

db.users.ensureIndex({ 
  "displayName": "text", 
  "about": "text", 
  "email": "text" 
}, { 
  "name": "users_full_text", 
  "default_language": "en",
  "language_override": "en" 
});  

... to which Mongo replies:

{
  "createdCollectionAutomatically" : false,
  "numIndexesBefore" : 3,
  "numIndexesAfter" : 4,
  "ok" : 1
}

Hurrah.

Answer 2

The value "en-US" is illegal as a language for text indexing according to doc page http://docs.mongodb.org/manual/reference/text-search-languages/#text-search-languages

MongoDB accepts only a 2 letter code or full name of language.

Hope it helps :)