DuckDB + Bruin

SourceDestination

Ingest data from DuckDB or push enriched data back, with quality checks, lineage, and scheduling. Defined in YAML, version-controlled in Git.

For business teams

What you get

Real-time warehouse sync
DuckDB tables replicate to your warehouse continuously. Analytics teams work with fresh data, not yesterday's export.
Catch issues at the source
Quality checks validate DuckDB data as it replicates. Null IDs, duplicate records, and schema drift get caught early.
Multi-source joins
Combine DuckDB with SaaS data, APIs, and other databases in your warehouse. One Bruin pipeline handles it all.
No untracked scripts
Replication is defined in YAML, reviewed in PRs, and deployed with CI/CD. No more mystery cron jobs.

For data & engineering teams

How it works

CDC with merge strategy
Bruin handles change data capture from DuckDB with deduplication. Schema changes are detected and handled automatically.
YAML-defined, Git-versioned
Your DuckDB replication is a YAML file. Review in PRs, deploy with CI/CD. No more untracked database scripts.
Row-level quality checks
Validate primary keys, foreign keys, and referential integrity on every sync. Catch corruption at the source.
Multi-source pipelines
Combine DuckDB with SaaS APIs and other databases in one pipeline. Bruin resolves cross-source dependencies.

Before you start

DuckDB installed locally or database file accessible

Write permissions for database file location

Sufficient memory for in-memory operations

Compatible file format version

Step 1

Add your DuckDB connection

DuckDB connection to local database file or in-memory. Add this to your Bruin environment file, credentials are stored securely and referenced by name in your pipeline YAML.

Parameters

pathPath to database file (use :memory: for in-memory)
read_onlyOptional flag for read-only access
threadsNumber of threads to use

connections:
  duckdb:
    type: duckdb
    uri: "duckdb:///path/to/database.duckdb"

Step 2

Create your pipeline

Define a YAML asset that tells Bruin what to pull from DuckDB and where to land it. This file lives in your Git repo, reviewable, version-controlled, and deployable with CI/CD.

Available tables

main.analyticsstaging.importsexports.reports

name: raw.duckdb_main.analytics
type: ingestr

parameters:
  source_connection: duckdb
  source_table: 'main.analytics'
  destination: bigquery

Step 3

Add quality checks

Add column-level and custom SQL checks to your DuckDB data. If a check fails, the pipeline stops, bad data never reaches downstream models or dashboards.

Validate row counts are within expected range

Ensure primary keys are unique and not null

Catch schema drift with freshness checks

columns:
  - name: id
    checks:
      - name: not_null
      - name: unique
  - name: created_at
    checks:
      - name: not_null

custom_checks:
  - name: row count within expected range
    query: |
      SELECT COUNT(*) BETWEEN 1 AND 10000000
      FROM raw.duckdb_main.analytics

Step 4

Run it

One command. Bruin connects to DuckDB, pulls data incrementally, runs your quality checks, and lands clean data in your warehouse. If a check fails, the pipeline stops, bad data never reaches downstream.

Backfill historical data with --start-date

Schedule with cron or trigger from CI/CD

Full lineage from DuckDB to your dashboards

$ bruin run .

Running pipeline...

  duckdb_main.analytics
    ✓ Fetched 2,847 new records
    ✓ Quality: campaign_id not_null     PASSED
    ✓ Quality: spend not_null           PASSED
    ✓ Quality: no negative ad spend     PASSED
    ✓ Loaded into bigquery

  Completed in 12s

Other Database integrations

Amazon DynamoDB

ClickHouse

Couchbase

CrateDB

Elasticsearch

Google Cloud Spanner

Ready to connect DuckDB?

Start for free, or book a demo to see how Bruin handles ingestion, quality, lineage, and scheduling for your entire data stack.

Book a demo Read the docs