find string containing special characters in mongodb using $regex

Question

using $regex in mongodb, I want to find the name B&B Hôtel which contain some special characters like & and ô by typing BB Hotel.

I tried this code:

db.txt.find({ "name": {'$regex': query, $options:'i'}})

where query can be BB Hotel.

You can only type `B&B H[ôo]tel`. Else, you'd have to normalize the text. — Wiktor Stribiżew, May 12 '17 at 16:31
@WiktorStribiżew would you explain please , i'm a really beginner — nassim, May 12 '17 at 16:38
@WiktorStribiżew yes work whene i search for "B&B H[ôo]tel" but how a simple user will find it ? — nassim, May 12 '17 at 16:42
user can search for "bb hotel" and he expect the result to be "B&B Hôtel" — nassim, May 12 '17 at 16:43
thanks @WiktorStribiżew for your answer, but can explain to me what is the basic of normalization of text — nassim, May 12 '17 at 16:45
Normalization means converting accented letters to their base representations. However, I see there might be a way to [set collations in MongoDB](https://derickrethans.nl/mongodb-collation-revised.html). — Wiktor Stribiżew, May 12 '17 at 16:50

Tezra · Answer 1 · 2017-05-12T17:07:05.210

1

You don't want regex search, you want diacritic insensitive text search

"name":{
  $text:
    {
      $search: "\"B&B Hotel\""
      $caseSensitive: false,
      $diacriticSensitive: false
    }
}

Note that $diacriticSensitive defaults to false, but I never trust the defaults. If you are running with older indexes (version 2 or less text index), you may not be able to use the indexes. The escaped " in the search part is to search for this phrase.

edited May 12 '17 at 17:07

answered May 12 '17 at 17:01

Tezra

8,463
3
31
68

It is a nice one, I could not find it. Do you know how exactly this works: does it apply the regular Unicode normalization (like with [NFC/NFD forms](http://stackoverflow.com/questions/16467479/normalizing-unicode)) or does it only deal with combining marks? I could not find these details [here](https://docs.mongodb.com/manual/reference/operator/query/text/#text-operator-diacritic-sensitivity). – Wiktor Stribiżew May 12 '17 at 17:12
@WiktorStribiżew By definition it works with unicode. The ^ above ô is a diacritic (and any letter not A-z is unicode). A diacritic is any mark on/above/below a character that changes the way it is meant to be pronounced. So your question, as you stated it, is how to do diacritic insensitive text search. – Tezra May 12 '17 at 17:17
@WiktorStribiżew Your other option is before plugging in "Hotel", do a "Hotel".replaceAll("[oô]","[oô]")... – Tezra May 12 '17 at 17:29

find string containing special characters in mongodb using $regex

1 Answers1