Database Selection

Choosing the right database for your Memori deployment.

Decision Matrix

Use Case	Database	Why
Development / prototyping	SQLite	Zero setup, file-based
Production (any scale)	PostgreSQL	Best concurrency, pooling, reliability
Existing MySQL infra	MySQL	No new database to manage
Enterprise Oracle shop	Oracle	Leverage existing licenses
Document-oriented needs	MongoDB	Flexible schema, horizontal scaling
CI / testing	SQLite	Fast, disposable, no dependencies

Feature	SQLite	PostgreSQL	MySQL	Oracle	MongoDB
Concurrent Writes	Limited	Excellent	Good	Excellent	Good
Connection Pooling	N/A	SQLAlchemy	SQLAlchemy	SQLAlchemy	Built-in
Cloud Hosting	N/A	Many	Many	Oracle Cloud	Atlas
Setup	None	Low	Low	Medium-High	Low
Driver	None	`psycopg2-binary`	`pymysql`	`oracledb`	`pymongo`

SQLite — Development, prototyping, single-user apps. One writer at a time — not for multi-user web apps.

PostgreSQL — Recommended for production. Concurrent reads/writes, connection pooling, available on every cloud platform.

MySQL — Good if your stack already runs MySQL. No reason to add another DB.

Oracle — Only if you already have Oracle infrastructure and DBA support.

MongoDB — Only NoSQL option. Uses PyMongo instead of SQLAlchemy but integrates fully with Memori.

Only the engine creation changes — all other Memori code stays the same:

Migration

engine = create_engine("sqlite:///memori.db")

After switching, run mem.config.storage.build() to create tables in the new database. Memori creates the same schema across all databases.