Firebase Database Issue in searching users by username initials and returning all users three

this is the way i store my data on Firebase DB:

     UUID : {
        username : "a"

     UUID :{
        username : "b"

I have already Indexed the username on rules. And this is the query for trying to retrieve data:

var dbRef = Database
        .database(url:"http//servername").reference(withPath: "users")
    dbRef.queryOrdered(byChild: "username").queryStarting(atValue: initials, childKey: "username")
        .observeSingleEvent(of: .value) { datasnap in
            var a = datasnap.value
            var b = a as! [String : Any]
            b.forEach { key, value in
                var a = value as! [String : Any]
                print(a["username"] as! String)

The problem that I have is :

  1. when I retrieve the dataSnapshot I have all the child three and not just the username
  2. the query results different strange values no matter of the initials

Does anybody have any solution or idea please, for actually searching usernames by their initials and not receiving random results(and also no case sensisitive).

There are two answers on stack for simmilar questions but are over 5-10 years ago and doesnt work. Thank you everybody,


  • This challenge is like knowhow - searching objects with string key for searching data on the firestore.

    let startTxt: String = searchKey
    let endTxt: String = searchKey + "\u{f8ff}"
        .whereField(User.key_uname, isGreaterThanOrEqualTo: startTxt)
        .whereField(User.key_uname, isLessThanOrEqualTo: endTxt)
        .limit(to: count)
        .getDocuments{ (doc, err) in ....

    Let me know if that works for you! :)

    It would be beneficial to save the usernames in lowercase in the database. This way, you can search for the usernames using updated start and end text.

    let startTxt: String = text.lowercased().trimmingCharacters(in: .whitespacesAndNewlines)
    let endTxt: String = (text + "\u{f8ff}").lowercased().trimmingCharacters(in: .whitespacesAndNewlines)

    This was working well with a database of 1,700,000 users.