A partial index includes only documents that satisfy a specified filter expression. It can reduce index size and focus the index on a frequently queried subset, such as active records or documents that have a particular state.

A sparse index omits documents that do not contain the indexed field. Sparse and partial indexes are not interchangeable: partial indexes support more expressive filtering criteria, subject to MongoDB's supported partial-filter expressions.

The query planner must be able to use the index without returning documents that should not match the query. Therefore, a query often needs to include conditions compatible with the partial filter before the partial index can be used safely.

Choose partial indexes when a subset is meaningful and common in the workload. Be cautious with sparse indexes when missing-field semantics are important.

db.users.createIndex(
  { emailNormalized: 1 },
  { partialFilterExpression: { active: true } }
);

// This query matches the partial filter.
db.users.find({ active: true, emailNormalized: "user@example.com" })
  .explain("executionStats");