Oracle → MongoDB

Why Migrate?

Strategic rationale, architecture shift

Developer Deep Dive

Data model, queries, transactions, HA

Migration Tooling

Relational Migrator, schema analysis, code gen

Feature Matrix & Battlecard

Oracle → MongoDB mapping, positioning

Data Model & Performance

Shift from normalization to rich documents designed around access patterns. Atomic single-document operations replace expensive multi-table JOINs.

Agility & Iteration Speed

Oracle's fixed DDL requires schema approval for every change. MongoDB's flexible schema lets developers iterate and deploy immediately.

Operational Simplicity

No DBA teams for patching, RMAN, capacity-based licensing. MongoDB Atlas automates scaling, backups, and upgrades.

Cloud & Vendor Flexibility

Oracle locks you to Exadata/OCI. Atlas runs on AWS, Azure, and GCP — multi-cloud by design, zero vendor lock-in.

AI Readiness

Store transactional data and vector embeddings in one platform. Atlas Vector Search enables RAG and semantic search — Oracle requires bolt-on AI infrastructure.

Licensing Cost

Oracle per-core licensing + mandatory support contracts + extra-cost options (RAC, partitioning, GoldenGate). MongoDB: pay-as-you-go consumption.

Oracle — Normalized Tables

-- 3 tables, 2 JOINs to fetch one order
SELECT o.order_id, c.name, p.product_name
FROM orders o
JOIN customers c ON o.customer_id = c.id
JOIN order_items oi ON o.id = oi.order_id
JOIN products p ON oi.product_id = p.id
WHERE o.order_id = 12345;

Separate tables for every entity. Foreign keys enforce relationships. Every read = multi-table JOIN.

MongoDB — Embedded Document

{
  "_id": "order_12345",
  "customer": { "name": "Jane Doe", "id": "C1" },
  "items": [
    { "product": "Hiking Boots", "qty": 1, "price": 129.99 },
    { "product": "Wool Socks", "qty": 2, "price": 14.99 }
  ],
  "total": 159.97
}

One document = one read. Atomic. No JOINs. Arrays and nested objects are first-class citizens.

Oracle SQL

-- Filter + Sort
SELECT * FROM users
WHERE status = 'active' AND age > 30
ORDER BY created_at DESC;

-- GROUP BY
SELECT dept, COUNT(*) FROM employees
GROUP BY dept;

MongoDB MQL

// Filter + Sort
db.users.find({
  "status": "active",
  "age": { $gt: 30 }
}).sort({ "created_at": -1 });

// Aggregation Pipeline (GROUP BY)
db.employees.aggregate([
  { $group: { _id: "$dept", count: { $sum: 1 } } }
]);

Oracle Transaction

-- Implicit transaction (auto-begin)
UPDATE accounts SET balance = balance - 100
  WHERE id = 'A';
UPDATE accounts SET balance = balance + 100
  WHERE id = 'B';
COMMIT; -- or ROLLBACK

MongoDB Multi-Document Transaction

const session = client.startSession();
session.startTransaction({
  readConcern: { level: "snapshot" },
  writeConcern: { w: "majority" }
});
try {
  accounts.updateOne(
    { _id: "A" }, { $inc: { balance: -100 } }, { session });
  accounts.updateOne(
    { _id: "B" }, { $inc: { balance: 100 } }, { session });
  session.commitTransaction();
} catch(e) { session.abortTransaction(); }

Oracle HA/DR

Data Guard: Standby databases (physical/logical). Manual or scripted failover. RPO depends on redo log shipping frequency.

GoldenGate: Real-time replication across heterogeneous systems. Complex setup, extra-cost license, dedicated infrastructure.

RAC: Multi-instance on shared storage. HA within one data center only. Does not protect against site failure.

MongoDB HA/DR

Replica Sets: 3+ nodes with automatic failover in seconds. No manual intervention. Built into every deployment.

Change Streams: Real-time event-driven architecture. Subscribe to the oplog — triggers, CDC, audit trail — no GoldenGate needed.

Global Clusters: Atlas deploys across continents with zone-aware sharding. Low-latency reads from the nearest node. RPO ≈ 0 with majority write concern.

Oracle Index Types

-- Composite index
CREATE INDEX idx_cust_date
  ON policies(customer_id, issue_date DESC);

-- Bitmap (low-cardinality columns)
CREATE BITMAP INDEX idx_status
  ON policies(status);

-- Function-based
CREATE INDEX idx_upper_name
  ON customers(UPPER(name));

MongoDB Index Types

// Compound (ESR: Equality, Sort, Range)
db.policies.createIndex(
  { customerId: 1, issueDate: -1 }
);

// Text index (full-text search)
db.products.createIndex({ description: "text" });

// Wildcard (dynamic schemas)
db.events.createIndex({ "metadata.$**": 1 });

// Vector (Atlas Vector Search)
// Defined in Atlas Search index definition

Schema Analysis

Reads Oracle data dictionary, foreign keys, and constraints. Recommends embedding vs. referencing based on relationship cardinality and access patterns.

Data Migration

Snapshot + CDC modes. Validates row counts between source and target. Handles type mapping (NUMBER → Double, DATE → ISODate). RM Docs →

Code Generation

Rewrites PL/SQL stored procedures to Java/Node.js/Python + Aggregation Pipelines. LLM-assisted for complex procedures. Breaks large blocks for context windows.