Q27: What is the RAM limit for an aggregation stage, and how do you bypass it?

Q21: What is the MongoDB Aggregation Framework and how does it work?

The Aggregation Framework is a data processing pipeline model. Documents enter a multi-stage pipeline that transforms, filters, groups, sorts, and reshapes them into aggregated results. Each stage passes its calculated output directly to the next stage.

Key Takeaways

Multi-stage transformation pipeline.
Executes native database analytics bypassing application memory processing.

Knowledge Check

How do aggregation stages pass data forward down the pipeline?

Q22: Explain $match, $group, and $project pipeline stages.

$match: Filters documents to allow only those matching specified conditions to pass (place $match first to leverage indexes).
$group: Groups input documents by a specified _id expression and applies accumulator operators ($sum, $avg, $push).
$project: Reshapes documents by adding, removing, or renaming fields.

Basic Aggregation Pipeline Example

db.orders.aggregate([
  { $match: { status: 'completed' } }, // 1. Filter
  { $group: {
      _id: '$customerId', 
      totalSpent: { $sum: '$amount' }, // 2. Group & Sum
      orderCount: { $sum: 1 }
  }},
  { $sort: { totalSpent: -1 } } // 3. Sort
]);

💡 Filters completed orders, groups by customer ID, calculates total expenditure, and orders results.

Key Takeaways

Place $match as early as possible to reduce document volume.
$group performs SQL GROUP BY aggregations.

Knowledge Check

Why should $match stages be placed at the very beginning of an aggregation pipeline?

Q23: How does the $lookup stage perform left outer joins between collections?

The $lookup stage performs a left outer join to pull documents from another collection into an array field:

from: Target collection name.
localField: Key field in current collection.
foreignField: Key field in target collection.
as: Output array field name.

Key Takeaways

Performs left outer joins between collections.
Outputs matching joined documents into an array field.

Knowledge Check

What data type does the $lookup stage produce for the joined output 'as' field?

Q24: What does the $unwind stage do in an aggregation pipeline?

The $unwind stage deconstructs an array field from input documents to output a document for each element of the array. Each output document replaces the array with an individual element item. Commonly paired with $lookup to flatten joined arrays.

Key Takeaways

Deconstructs array fields into individual document streams.
Use preserveNullAndEmptyArrays: true to keep documents missing the array.

Knowledge Check

If a document contains an array of 3 items, how many output documents will $unwind produce for that document?

Q25: What are $facet and $bucket stages used for?

$facet: Executes multiple aggregation sub-pipelines concurrently within a single stage on the same input documents (useful for multi-faceted search and dashboard filtering).
$bucket: Categorizes incoming documents into groups (buckets) based on specified boundaries and criteria.

Key Takeaways

$facet powers multi-faceted search breakdowns in a single query.
$bucket categorizes numbers into defined numerical range groups.

Knowledge Check

Which stage allows running parallel aggregation pipelines to compute search results and category totals in one query pass?

Q26: What is the purpose of $addFields (or $set) in aggregation?

The $addFields stage adds new computed fields or modifies existing fields in passing documents without removing un-mentioned existing fields (unlike $project, which excludes un-specified fields unless explicitly preserved).

Key Takeaways

Adds or overrides specific document fields.
Preserves all existing document keys automatically.

Knowledge Check

How does $addFields differ from $project in aggregation?

Pipeline stages have a default memory limit of 100 Megabytes of RAM. If a stage (like $sort or $group) exceeds 100MB, MongoDB throws an error. To bypass this, pass { allowDiskUse: true }, allowing pipeline stages to spill temporary data to disk files.

Key Takeaways

Default aggregation stage RAM ceiling: 100MB.
Use { allowDiskUse: true } to write intermediate data to temporary disk space.

Knowledge Check

Which query option allows aggregation stages to exceed the 100MB RAM processing limit?

Q28: How does Mongoose .populate() work, and what are its performance trade-offs?

populate() automates replacing specified document path references (ObjectId) with actual documents from another collection. Trade-off: Under the hood, populate() executes separate additional queries in application code (not a native database engine join), which can lead to $N+1$ query performance issues if misused.

Mongoose Populate Usage

const post = await Post.findById(postId)
  .populate('author', 'name email') // Replace author ObjectId with name and email
  .exec();

💡 Retrieves post documents and executes a secondary query to fetch matching author details.

Key Takeaways

Fills referenced ObjectIds with full document objects.
Executes secondary queries under the hood; use $lookup for complex native joins.

Knowledge Check

Does Mongoose .populate() execute a native database join inside MongoDB engine memory?

Q29: What is the .lean() method in Mongoose, and why does it boost read performance?

By default, Mongoose query results return heavy Mongoose Document instances complete with internal state, getters/setters, change tracking, and method bindings. Chaining .lean() tells Mongoose to skip document hydration and return plain JavaScript objects (POJOs), making reads up to 5x faster.

Key Takeaways

Returns plain JavaScript objects instead of heavy Mongoose Document instances.
Drastically reduces CPU and memory overhead for read-heavy APIs.

Knowledge Check

What capability do you lose when executing a query with .lean() in Mongoose?

Q30: Does MongoDB support ACID Transactions? How do multi-document transactions work?

Yes, MongoDB supports full multi-document ACID transactions (starting in version 4.0 for Replica Sets and 4.2 for Sharded Clusters). Transactions allow executing multiple read/write operations atomically using a session (session.startTransaction()). If an operation fails, calling abortTransaction() rolls back all updates.

Mongoose Multi-Document Transaction

const session = await mongoose.startSession();
session.startTransaction();
try {
  await Account.updateOne({ _id: fromId }, { $inc: { balance: -100 } }, { session });
  await Account.updateOne({ _id: toId }, { $inc: { balance: 100 } }, { session });
  await session.commitTransaction();
} catch (err) {
  await session.abortTransaction();
} finally {
  session.endSession();
}

💡 Executes account balance updates inside a session transaction, committing or aborting atomically.

Key Takeaways

Multi-document ACID transactions require Replica Sets or Sharded Clusters.
Managed via Client Sessions (session.commitTransaction()).

Knowledge Check

Which deployment architecture is required to run multi-document ACID transactions in MongoDB?

70 Progressive Questions (Easy to Advanced + Scenarios)

Q21: What is the MongoDB Aggregation Framework and how does it work?

Q22: Explain $match, $group, and $project pipeline stages.

Q23: How does the $lookup stage perform left outer joins between collections?

Q24: What does the $unwind stage do in an aggregation pipeline?

Q25: What are $facet and $bucket stages used for?

Q26: What is the purpose of $addFields (or $set) in aggregation?