MongoDB Query Builder

find() returns a cursor of all matching documents — use it to fetch multiple records with optional sort, limit, and skip. findOne() is equivalent to find().limit(1) and returns a single document object (or null) instead of a cursor — ideal when you know only one record matches. aggregate() uses a pipeline of stages ($match, $group, $project, $sort, $lookup, etc.) to transform documents; it's more powerful than find() for computations, joins, grouping, and reshaping data. For simple CRUD reads, prefer find/findOne; use aggregate when you need computed fields, joined data, or grouped results.

Projection controls which fields are returned in each document. Pass an object where 1 includes a field and 0 excludes it: { name: 1, email: 1 } returns only name and email (plus _id by default). You cannot mix inclusions and exclusions in the same projection — except for _id, which you can explicitly exclude alongside inclusions: { name: 1, _id: 0 }. Projections reduce network transfer and memory usage when documents are large but you only need a subset of fields.

$eq is the implicit default — db.users.find({ age: 25 }) is shorthand for { age: { $eq: 25 } }. Comparison operators ($gt, $gte, $lt, $lte, $ne) work on numbers, dates, and strings. $in and $nin match against an array of possible values: { status: { $in: ['active', 'pending'] } }. $regex matches string fields against a regular expression. $exists checks whether a field is present: { phone: { $exists: true } }. Combine multiple operators with $and (implicit when using multiple fields) or $or for alternative conditions.

The native MongoDB driver (@mongodb/mongodb) gives you direct access to all MongoDB operations with minimal abstraction — you work with plain JavaScript objects and the full query API. Mongoose adds a schema layer on top: you define models with types, validation rules, and default values, and it enforces those at the application level. Mongoose also provides middleware (pre/post hooks), virtual fields, and helper methods. For simple apps or microservices where flexibility matters more than schema enforcement, the native driver is lighter. For full-stack apps that benefit from data validation and model organization, Mongoose is the popular choice.

Use aggregate when you need any of the following: $group to count or sum values (e.g., total orders per user), $lookup to join with another collection (like SQL JOIN), $unwind to flatten arrays into individual documents, $addFields or $project to compute new fields (e.g., concatenate name parts, calculate age from birthdate), or multi-stage data transformation pipelines. For simple filtering, sorting, and pagination without computed fields or joins, find() with sort/limit is faster and simpler. MongoDB also optimizes find() with index usage more aggressively than complex aggregate pipelines.