Search Shortcut cmd + k | ctrl + k
mongo
Downloads 2.0kthis week
GitHub stars 57
Extension descriptor (YAML)

Integrates DuckDB with MongoDB, enabling direct SQL queries and writes over MongoDB collections without ETL

Maintainer(s): stephaniewang526

Installing and Loading 

INSTALL mongo FROM community;
LOAD mongo;

Example 

-- Attach to MongoDB
ATTACH 'host=localhost port=27017' AS mongo_db (TYPE MONGO);

-- Query collections with SQL
SELECT * FROM mongo_db.mydb.mycollection LIMIT 10;

-- Load a CSV directly into MongoDB
INSERT INTO mongo_db.mydb.mycollection
SELECT * FROM read_csv('data.csv');

About mongo

The duckdb-mongo extension provides direct SQL access to MongoDB collections without requiring data export or ETL processes. It supports both standalone MongoDB instances and MongoDB Atlas clusters.

Reading data:

  • Automatic schema inference from sampled documents, with nested document flattening
  • User-provided schemas and Atlas SQL __schema document support
  • Query pushdown to MongoDB for filters, projections, limits, TopN, and aggregations (COUNT, SUM, MIN, MAX, AVG)
  • BSON type mapping including ObjectId, Decimal128, arrays, and nested documents

Writing data (new in v0.3.0):

  • INSERT INTO – insert rows from any DuckDB source (CSV, Parquet, other databases, subqueries)
  • CREATE TABLE AS SELECT – create new MongoDB collections from query results
  • COPY TO (FORMAT mongo) – bulk-load data into MongoDB collections
  • DROP TABLE – drop MongoDB collections

Connections and security:

  • MongoDB Atlas support via connection strings or DuckDB Secrets
  • TLS/SSL encryption, custom CA certificates
  • mongo_enable_direct_scan setting to restrict mongo_scan to attached catalogs only

For detailed setup and usage instructions, visit the extension repository.

Added Functions

function_name function_type description comment examples
mongo_clear_cache table Clears the schema cache for all attached MongoDB databases. Useful when MongoDB schema changes. NULL [SELECT * FROM mongo_clear_cache()]
mongo_scan table Scans a MongoDB collection and returns its contents as a table. Supports optional filter and sample_size parameters. NULL [SELECT * FROM mongo_scan('mongodb://localhost:27017', 'mydb', 'mycollection'), SELECT * FROM mongo_scan('mongodb://localhost:27017', 'mydb', 'mycollection', filter := '{"status": "active"}')]

Overloaded Functions

This extension does not add any function overloads.

Added Types

This extension does not add any types.

Added Settings

name description input_type scope aliases
mongo_enable_direct_scan Allow mongo_scan to use a connection string that is not from an attached MongoDB database. When false, mongo_scan is only allowed if the connection string exactly matches an attached Mongo catalog. BOOLEAN GLOBAL []