MongoDB Cheat Sheet

Updated 2026-07-17

🎯Take a practice test on this topic17 practice tests · 547 questions→

🧠Study flashcards on this topic185 cards · spaced repetition→

MongoDB is an open-source, document-oriented NoSQL database that stores data in flexible, JSON-like BSON documents. Known for horizontal scalability, rich querying, and a powerful aggregation framework, MongoDB is used by organizations ranging from startups to Fortune 500 companies. The current stable release is the MongoDB 8.3 series, continuing the 8.x line that opened with 8.0 (GA October 2024) and brought a bulkWrite command spanning multiple collections, updateOne with sort, Queryable Encryption range queries, automatic compaction, per-CPU TCMalloc caches, and significant performance improvements. MongoDB Atlas, the fully managed cloud service, continues to expand with Atlas Vector Search, Atlas Stream Processing, and Atlas Search. This cheat sheet covers the complete MongoDB ecosystem from basic CRUD operations and query operators through advanced aggregation, indexing strategies, replication, sharding, security, and Atlas-specific features.

What This Cheat Sheet Covers

This topic spans 51 focused tables and 419 indexed concepts, 185 flashcards, 17 practice tests with 547 questions. Below is a complete table-by-table outline of this topic, spanning foundational concepts through advanced details.

Table 1: BSON Data TypesTable 2: CRUD Operations — InsertTable 3: CRUD Operations — Read (Query)Table 4: CRUD Operations — UpdateTable 5: CRUD Operations — DeleteTable 6: Comparison Query OperatorsTable 7: Logical Query OperatorsTable 8: Element Query OperatorsTable 9: Evaluation Query OperatorsTable 10: Array Query OperatorsTable 11: Bitwise Query OperatorsTable 12: Update Operators — FieldTable 13: Update Operators — ArrayTable 14: Aggregation Pipeline — Core StagesTable 15: Aggregation Pipeline — Advanced StagesTable 16: Aggregation Pipeline — Atlas and Diagnostic StagesTable 17: Aggregation AccumulatorsTable 18: Aggregation — Conditional and Utility OperatorsTable 19: Aggregation — Date OperatorsTable 20: Aggregation — String OperatorsTable 21: Aggregation — Array and Set OperatorsTable 22: Aggregation — Type and Conversion OperatorsTable 23: Aggregation — Math OperatorsTable 24: Window Function OperatorsTable 25: Index TypesTable 26: Index ManagementTable 27: Query OptimizationTable 28: Projection OperatorsTable 29: Cursor MethodsTable 30: Multi-Document TransactionsTable 31: ReplicationTable 32: Read PreferenceTable 33: ShardingTable 34: Write Concern LevelsTable 35: Bulk OperationsTable 36: Collection OperationsTable 37: Capped CollectionsTable 38: Schema ValidationTable 39: Time Series CollectionsTable 40: GridFSTable 41: Change StreamsTable 42: Text SearchTable 43: Geospatial QueriesTable 44: CollationTable 45: Views and On-Demand Materialized ViewsTable 46: Database CommandsTable 47: Security and AuthenticationTable 48: Backup and RestoreTable 49: MongoDB Atlas FeaturesTable 50: Performance Monitoring and ProfilingTable 51: Connection String and Pooling

Quick IndexSubscribe to unlock

A jump-to index of every table row in this cheat sheet.

Mind MapSubscribe to unlock

An interactive map of every table and concept in this topic.

Table 1: BSON Data Types

MongoDB stores every document as BSON, a binary format carrying real types that JSON never had, such as Date, ObjectId, Binary and Decimal128. Choosing the right one matters more than it looks: Decimal128 keeps money exact where Double quietly rounds it, Timestamp is internal replication plumbing rather than a way to store dates, and _id accepts far more than just an ObjectId. These are the primitives every document is built from.

BSON Type	Example	Description
Double	`{ rating: 4.5 }`	• 64-bit binary floating point (IEEE 754) • approximates decimal fractions, so avoid it for money
String	`{ name: "MongoDB" }`	UTF-8 encoded character string
Object	`{ addr: { city: "NYC" } }`	• Embedded/nested BSON document • fields are ordered, so whole-document equality is order-sensitive
Array	`{ tags: ["db", "nosql"] }`	• Ordered list of values of any BSON type • an ascending sort compares each array's smallest element
Binary Data	`BinData(0, "data")`	• Byte array plus a subtype that says how to interpret it • used for files, encrypted fields, UUIDs
ObjectId	`ObjectId("507f1f77bcf86cd799439011")`	• 12 bytes: 4-byte creation timestamp, 5-byte random, 3-byte counter • the `_id` a driver generates when you omit one • roughly, not strictly, ordered by creation time
Boolean	`{ active: true }`	• `true` or `false` value • a distinct type from the number `1`, matched by `$type: "bool"`
Date	`{ created: ISODate("2026-04-21") }`	• Signed 64-bit integer, milliseconds since the Unix epoch • build it with `new Date()` or `ISODate()`, since bare `Date()` returns a string
Null	`{ middle: null }`	• An explicit null value, not the same thing as an absent field • sorts below every number
Regular Expression	`{ pattern: /^mongo/i }`	• Perl-compatible pattern, PCRE2 since MongoDB 6.1 • storable as a field value, not only usable in a query
32-bit Integer	`{ count: Int32(42) }`	• 32-bit signed integer • `Int32()` in mongosh, legacy alias `NumberInt()` • what a bare whole number becomes in mongosh
Timestamp	`Timestamp(1714700000, 1)`	• Internal type for the replication oplog • not for date storage, use `Date` instead
64-bit Integer	`{ pop: Long("7800000") }`	• 64-bit signed integer • `Long()` in mongosh, legacy alias `NumberLong()` • a bare number too big for `Int32` becomes a `Double`, not a `Long`
Decimal128	`{ amt: Decimal128("9.99") }`	• 128-bit decimal floating point, 34 digits of precision • exact decimal values, the recommended type for money
UUID	`UUID("3b241101-e2bb-4255-8caf-4136c566a962")`	• Stored as Binary subtype 4 • 16 bytes as BinData against 36 as a string, keeping the `_id` index smaller
MinKey / MaxKey	`{ lo: MinKey(), hi: MaxKey() }`	• Sentinels that compare lower/higher than every other type • exist primarily for internal use

Table 2: CRUD Operations — Insert

The first half of CRUD: getting documents into a collection. Both insert methods build the collection (and the database) on the first write and generate an _id when you leave one out. The two options are where bulk loads get interesting: ordered: false keeps a batch going past its first bad document, and writeConcern raises durability for a single write without touching the rest.

Operation	Example	Description
insertOne()	`db.users.insertOne({ name: "Alice", age: 30 })`	• Inserts a single document • Returns `acknowledged` plus `insertedId`, never the document itself

The rest of this section is part of the full cheat sheet

Table 3: CRUD Operations — Read (Query)

Reading data back. find() returns a cursor and findOne() a single document, while projections trim the fields you get back. Watch the counting pair: countDocuments() is accurate but scans, whereas estimatedDocumentCount() is instant but only reports the whole collection from metadata.

Operation	Example	Description
find()	`db.users.find({ age: { $gt: 25 } })`	Returns a cursor to the matching documents, not the documents themselves

The rest of this section is part of the full cheat sheet

Table 4: CRUD Operations — Update

Modifying existing documents, with more nuance than insert or delete. The key distinctions: update operators like $set change individual fields while replaceOne swaps the whole document and drops anything you leave out; the findOneAndUpdate family hands back the before-or-after copy; upsert creates the document when nothing matched, and without it a non-matching update is a silent no-op; and an aggregation-pipeline update lets one field be computed from others.

Operation	Example	Description
updateOne()	`db.users.updateOne({ name: "Alice" }, { $set: { age: 31 } })`	• Updates at most one document, the first the filter matches • the update must be operator expressions or a pipeline, never a plain document • `sort` option (8.0+) chooses which match counts as first

The rest of this section is part of the full cheat sheet

Table 1: BSON Data Types

BSON Type	Example	Description
Double	`{ rating: 4.5 }`	• 64-bit binary floating point (IEEE 754) • approximates decimal fractions, so avoid it for money
String	`{ name: "MongoDB" }`	UTF-8 encoded character string
Object	`{ addr: { city: "NYC" } }`	• Embedded/nested BSON document • fields are ordered, so whole-document equality is order-sensitive
Array	`{ tags: ["db", "nosql"] }`	• Ordered list of values of any BSON type • an ascending sort compares each array's smallest element
Binary Data	`BinData(0, "data")`	• Byte array plus a subtype that says how to interpret it • used for files, encrypted fields, UUIDs
ObjectId	`ObjectId("507f1f77bcf86cd799439011")`	• 12 bytes: 4-byte creation timestamp, 5-byte random, 3-byte counter • the `_id` a driver generates when you omit one • roughly, not strictly, ordered by creation time
Boolean	`{ active: true }`	• `true` or `false` value • a distinct type from the number `1`, matched by `$type: "bool"`
Date	`{ created: ISODate("2026-04-21") }`	• Signed 64-bit integer, milliseconds since the Unix epoch • build it with `new Date()` or `ISODate()`, since bare `Date()` returns a string
Null	`{ middle: null }`	• An explicit null value, not the same thing as an absent field • sorts below every number
Regular Expression	`{ pattern: /^mongo/i }`	• Perl-compatible pattern, PCRE2 since MongoDB 6.1 • storable as a field value, not only usable in a query
32-bit Integer	`{ count: Int32(42) }`	• 32-bit signed integer • `Int32()` in mongosh, legacy alias `NumberInt()` • what a bare whole number becomes in mongosh
Timestamp	`Timestamp(1714700000, 1)`	• Internal type for the replication oplog • not for date storage, use `Date` instead
64-bit Integer	`{ pop: Long("7800000") }`	• 64-bit signed integer • `Long()` in mongosh, legacy alias `NumberLong()` • a bare number too big for `Int32` becomes a `Double`, not a `Long`
Decimal128	`{ amt: Decimal128("9.99") }`	• 128-bit decimal floating point, 34 digits of precision • exact decimal values, the recommended type for money
UUID	`UUID("3b241101-e2bb-4255-8caf-4136c566a962")`	• Stored as Binary subtype 4 • 16 bytes as BinData against 36 as a string, keeping the `_id` index smaller
MinKey / MaxKey	`{ lo: MinKey(), hi: MaxKey() }`	• Sentinels that compare lower/higher than every other type • exist primarily for internal use

Table 2: CRUD Operations — Insert

Operation	Example	Description
insertOne()	`db.users.insertOne({ name: "Alice", age: 30 })`	• Inserts a single document • Returns `acknowledged` plus `insertedId`, never the document itself

Table 3: CRUD Operations — Read (Query)

Operation	Example	Description
find()	`db.users.find({ age: { $gt: 25 } })`	Returns a cursor to the matching documents, not the documents themselves

Table 4: CRUD Operations — Update

Operation	Example	Description
updateOne()	`db.users.updateOne({ name: "Alice" }, { $set: { age: 31 } })`	• Updates at most one document, the first the filter matches • the update must be operator expressions or a pipeline, never a plain document • `sort` option (8.0+) chooses which match counts as first