Using $in clause for insensitive search in mongo db driver c#

Question

I am wondering how to use the case insensitive for $in expressions.

According to the official MongoDB manual you can do this:

{ name: { $in: [ /^acme/i, /^ack/ ] } }

I tested out this on Compass and it's working fine, search is insensitive.

I need to this using the Mongo Driver in C#.

I am doing this:

  var array = new BsonArray(companyNames);

  var filter = new BsonDocument { { "Name", new BsonDocument { { "$in", new BsonArray(array) }} } };
  var result = _collection.Find(filter).ToList();

companyNames is a string[]

However this retrieves me only exact matches. It's kind of obvious because I am not including the "regex" expression. But I don't know how I am able to include the regex in a string.

The workaround is to create an $or expresion with regex for every company name.

Does anyone knows how to do this?

Thanks

Answer 1

With mongo-csharp-driver , you can utilise MongoDB.Bson.BsonRegularExpression . You can either perform:

var arrayIn = new BsonArray().Add(
                      new BsonRegularExpression("^acme", "i")
                  ).Add(
                      new BsonRegularExpression("^ack"));
var filter = new BsonDocument { { "name", new BsonDocument { { "$in", arrayIn }} } };
var cursor = collection.Find(filter).ToList();

Or alternatively, instead of string use Regex and RegexOptions :

var arrayIn = new BsonArray().Add(
                      new BsonRegularExpression(
                          new Regex(
                              "^acme", RegexOptions.IgnoreCase))
                  ).Add(new BsonRegularExpression(
                              "^ack"));

Answer 2

here's an elegant solution using a text index on the company name field.

using MongoDB.Entities;

namespace StackOverflow
{
    class Program
    {
        public class Company : Entity
        {
            public string Name { get; set; }
        }

        static void Main(string[] args)
        {
            new DB("test");

            DB.Index<Company>()
              .Key(c => c.Name, KeyType.Text)
              .Option(o => o.Background = false)
              .Create();

            var c1 = new Company { Name = "Ackme" };
            var c2 = new Company { Name = "Acme" };
            var c3 = new Company { Name = "Lackme" };
            var c4 = new Company { Name = "Hackme" };

            c1.Save(); c2.Save(); c3.Save(); c4.Save();

            var names = new[] { "ackme", "acme" };

            var result = DB.SearchText<Company>(string.Join(" ", names));
        }
    }
}

mind you, the above uses the convenience library MongoDB.Entities . however the concepts are the same but the syntax for the official driver is cumbersome compared to the above.

Answer 3

My solution was an overloaded .In() as an extension method, taking an extra parameter for case-insensitivity:

public static class MongoExtensions
{
    public static FilterDefinition<TDocument> In<TDocument>(
        this FilterDefinitionBuilder<TDocument> builder,
        Expression<Func<TDocument, object>> expr,
        IEnumerable<string> values,
        bool ignoreCase)
    {
        if (!ignoreCase)
        {
            return builder.In(expr, values);
        }

        var filters = values
            .Select(v => builder.Regex(expr, new BsonRegularExpression($"^{Regex.Escape(v)}$", "i")));

        return builder.Or(filters);
    }
}

Then you can simply:

var filter = Builders<Companies>.Filter.In(c => c.Name, companyNames, true);
var result = _collection.Find(filter).ToList();