Find the fast path.

Create a useful index

const users = BsonLite.collection('users');
await users.createIndex({ email: 1 }, { unique: true });
await users.createIndex({ active: 1, createdAt: -1 });

const indexes = await users.listIndexes().toArray();

A unique index rejects duplicate indexed values. Existing duplicates can prevent index creation. Compound field order matters: choose indexes for your filters and sort order, then validate with real data.

Text search

const notes = BsonLite.collection('notes');
await notes.createIndex({ title: 'text', body: 'text' });

const matches = await notes.find({
  $text: { $search: 'portable database' }
}).toArray();

Text search uses the server capabilities and language behavior of MongoDB 4.0. It requires a text index. It is different from Atlas Search, which this local SDK does not provide.

Keep index costs visible

• Each index consumes disk space and adds work to writes.
• Index construction can take time on existing collections.
• Use supported MongoDB 4.0 index options. Native encryption is enforced after collection/index storage settings.
• Drop an index only after checking the application queries that rely on it.

// Example: drop an intentionally retired non-_id index.
await users.dropIndex('active_1_createdAt_-1');