MongoDB Developer Associate study plan

Set up a local MongoDB 7.x instance and mongosh, then create two databases with three collections each and insert at least 50 documents total using insertOne, insertMany, and ordered/unordered writes.

Write 15 queries that combine equality filters, comparison operators, and logical operators, then rewrite each one using dot notation on embedded documents and arrays where applicable.

Practice projection, sort, limit, skip, and countDocuments on your own dataset, and explain in writing when a query can be answered from an index versus a collection scan.

Build a one-page cheat sheet mapping CRUD method names to their shell equivalents and to driver equivalents in your chosen language (C#, Java, Node.js, PHP, or Python).

Design and create single-field, compound, multikey, unique, partial, and TTL indexes on your dataset, and capture explain("executionStats") output showing IXSCAN versus COLLSCAN.

Determine the winning plan for five queries and adjust index key order to satisfy ESR (equality, sort, range) rules, documenting before/after metrics.

Implement a covered query and prove it with explain output showing totalDocsExamined equal to 0.

Use aggregation stages match, project, group, sort, unwind, lookup, and set to answer five analytical questions about your dataset.

Write five aggregation pipelines that include at least one $lookup across collections and one $facet or $bucket stage, and verify results against hand-computed expectations.

Configure a three-node replica set locally, then demonstrate write concern majority and read concern majority/local behavior with a scripted failover test.

Implement optimistic concurrency using a version field and compareAndSwap-style updates with findOneAndUpdate, and record how many retries occurred under simulated contention.

Describe transaction semantics by running a multi-document transaction that aborts on error and confirming no partial writes remain.

Implement schema validation with $jsonSchema on two collections and verify that invalid inserts are rejected with the expected error codes.

Set up authentication with SCRAM, create roles with least privilege, and prove that a restricted user cannot perform an unauthorized operation.

Diagram an embedding-versus-referencing decision for three entities in a sample app and justify each choice using access patterns and document size limits.

Diagnose three deliberately introduced problems (missing index, unbounded array growth, oversized document) and document the fix for each.

Install the official driver for your chosen language and connect using both a direct connection string and a replica set connection string with retryWrites enabled.

Implement full CRUD for two collections using the driver's typed models or document classes, including bulk writes and upserts.

Implement cursor iteration with batching, and demonstrate correct resource cleanup and error handling for duplicate key and network timeout errors.

Build a small repository layer that isolates driver calls behind interfaces so query logic is testable without a live server.

Build aggregation pipelines in driver code using the driver's builder APIs rather than raw JSON strings, and compare readability against string pipelines.

Implement a multi-document transaction in the driver, including session passing and commit/abort handling with a forced failure case.

Add connection pool configuration, server selection timeout, and a retryable write scenario, then measure behavior when the primary is stepped down.

Write integration tests using an in-memory or containerized MongoDB instance and wire them into a single test command.

Complete two timed 60-question practice sets weighted 50/50 across core and language-specific material, and log every missed question by topic.

Re-derive answers for each missed question from official documentation, writing the doc section that confirms the correct behavior.

Re-implement the three weakest core topics (for example index selection, aggregation stage behavior, replication concerns) as small runnable demos.

Profile your per-topic accuracy and rank the five topics with the lowest scores for targeted review.

Review the error log and rewrite each weak topic as a one-paragraph explanation you can state without notes.

Complete two full-length timed mock exams under exam conditions and score at or above your target threshold on both.

Refactor your week 5-6 driver project to fix any anti-patterns you discovered during review, such as missing indexes or unbounded queries.

Run a final 30-question rapid drill on high-frequency topics and confirm no topic remains below 70 percent accuracy.