What is difference between partial indexes and sparse indexes mongodb?

Question

I've read official docs of MongoDB but really can't understand the difference between sparse and partial indexes. I wanted to have an explanatory view with examples.

Answer 1

Sparse index is an optimized index which only contains pointers to documents that have value(s) in the indexed fields. For example, let's say you would like to add an index on lastname field

{ _id: 1, firstname: 'John', lastname: 'Black', age: 20 }
{ _id: 2, firstname: 'Stive', lastname: 'White', age: 17 }
{ _id: 3, firstname: 'Tom', age: 22 }

if you run

db.users.createIndex({ lastname: 1 });

command, it will add indexes on 3 documents, but you don't need to have an index on a document where is no lastname value (_id: 3); it's a waste of space and memory. To avoid empty fields' indexing, mongodb has sparse index, which is simply "check for non-empty value". So when you add sparse: true

db.users.createIndex({ lastname: 1, sparse: true });

Mongodb will add indexes only for 2 documents (_id: 1, _id:2). Its great, but what if you want to index only those users' documents which are older than 18 years? You cant use sparse index because it only checks documents for value existence.

This is why partial indexes were created.

db.person.createIndex( 
  { age: 1}, 
  { partialFilterExpression: { age: { $gte: 18 }, lastname: { $exists: true }}
);

This example will put index only for 1 document(id: 1). Partial index is complex version of sparse, it will filter documents not only checking their existence, but using conditions provided in partialFilterExpression field.