Skip to content
Mauricio David edited this page Jan 31, 2020 · 22 revisions

This documentation is valid only for v4.x version.


LiteDB improves search performance by using indexes on document fields. Each index stores the value of a specific field ordered by the value (and type) of the field. Without an index, LiteDB must execute a query using a full document scan. Full document scans are inefficient because LiteDB must deserialize all documents to test each one by one.

Index Implementation

LiteDB uses a simple index solution: Skip Lists. Skip lists are double linked sorted list with up to 32 levels. Skip lists are super easy to implement (only 15 lines of code) and statistically balanced. The results are great: insert and find results has an average of O(ln n) = 1 million of documents = 13 steps. If you want to know more about skip lists, see this great video.

Documents are schema-less, even if they are in the same collection. So, you can create an index on a field that can be one type in one document and another type in another document. When you have a field with different types, LiteDB compares only the types. Each type has an order:

BSON Type Order
MinValue 1
Null 2
Int32, Int64, Double, Decimal 3
String 4
Document 5
Array 6
Binary 7
ObjectId 8
Guid 9
Boolean 10
DateTime 11
MaxValue 12
  • Numbers (Int32, Int64, Double or Decimal) have the same order. If you mix these number types in the same document field, LiteDB will convert them to Decimal when comparing.

Primary key (= auto id)

Primary key is also one of the indexes. By default primary key _id will be created and inserted automatically when you call method col.Insert() . However it won't be created if Custom types are used in Id property.

EnsureIndex()

Indexes are created via EnsureIndex. This instance method ensures an index: creates the index if it does not exist or does nothing if already exists. In v4 the indexes are not re-created automatically anymore when the definition changes. In case you would change the definition of an index, you have to drop it first, then call EnsureIndex again.

Indexes are identified by document field name. LiteDB supports only 1 field per index, but this field can be any BSON type, even an embedded document.

{
    _id: 1,
    Address:
    {
        Street: "Av. Protasio Alves, 1331",
        City: "Porto Alegre",
        Country: "Brazil"
    }
}
  • You can use EnsureIndex("Address") to create an index to all Address embedded document
  • Or EnsureIndex("Address.Street") to create an index on Street using dotted notation
  • Indexes are executed as BsonDocument fields. If you are using a custom ResolvePropertyName or [BsonField] attribute, you must use your document field name and not the property name. See Object Mapping.
  • You can use a lambda expression to define an index field in a strongly typed collection: EnsureIndex(x => x.Name)

MultiKey Index

When you create an index in an array type field, all values are included on index keys and you can search for any value.

public class Customer
{
    public int Id { get; set; }
    public string Name { get; set; }
    public string[] Phones { get; set; }
}

var customers = db.GetCollection<Customer>("customers");

customers.Insert(new Customer { Name = "John", Phones = new string[] { "1", "2", "5" });
customers.Insert(new Customer { Name = "Doe", Phones = new string[] { "1", "8" });

customers.EnsureIndex(x => x.Phones, "$.Phones[*]");

var result = customers.Query(x => x.Phones.Contains("1")); // returns both documents

Expressions

In v4 it's possible to create an index based on a expression execution with multikey values support. With this, you can index any king information that are not direct you field value, like:

  • db.EnsureIndex("customer", "Name", false, "LOWER($.Name)")
  • db.EnsureIndex("customer", "Total", false, "SUM($.Items[*].Price)")
  • db.EnsureIndex("customer", "CheapBooks", false, "LOWER($.Books[@.Price < 20].Title)")

See Expressions for more details about expressions.

Changes in v4

  • There is no more auto index creation. You always run EnsureIndex in your database initialization.
  • If you try query without an index, query will be runned using full search
  • If you are using a LINQ expression with no resolution to Query object, query engine will run query after map to object
  • When your query has an And operation, engine will run only 1 side with index (if exist) and another side will use full scan. This optimize results avoiding multi index queries. Try always use left
col.EnsureIndex(x => x.Name);
col.EnsureIndex(x= > x.Age);

var r = col.Find(x => x.Name == "John" && x.Age > 20 && x.Phones.Length > 1);

In this example, LiteDB will use Name index to get first results. For Age > 20 full scan will be used over all documents which Name == 'John'. And them, over result of this, query Phones.Length > 1

Limitations

  • Index values must have less than 512 bytes (after BSON serialization)
  • Max of 16 indexes per collections - including the _id primary key