Tangents
Federated Query Lake
Tangents empowers you to seamlessly query across multiple databases through a unified, no-code interface in SyneHQ. With just a few clicks, you can analyze, join, and visualize data from any supported source — all in one place.
How to Create and Use a Tangent
1. Go to the Tangents Page Navigate to data.synehq.com/tangents.
2. Create a Tangent
Click the Create Tangent button.
Name your Tangent.
Select the supported database connections (set these up at data.synehq.com/connections beforehand if needed).
3. Browse and Query
After creation, your Tangent appears as a card.
Click Browse & Query on the Tangent card to reach the data browser.
Here, you can:
Explore all tables across your connected databases
Visualize schema relationships
Use the Query Console (data.synehq.com/console) to run federated SQL
4. Managing Connections
Add or update your data sources at any time on the Connections page.
How Does Tangents Work Under the Hood?
Tangents leverages the DuckDB query engine to provide its federated capabilities. Here’s an inside look at the technology powering your experience:
DuckDB and the ATTACH Command
DuckDB is a high-performance, in-process analytics database optimized for interactive exploration and cross-database operations. A key feature used by Tangents is the ATTACH statement, which allows DuckDB to connect ("attach") to multiple databases simultaneously—even databases from different systems like PostgreSQL, MySQL, SQLite, or additional DuckDB files.duckdb.org
What Does This Mean for You?
Query Multiple Databases at Once: DuckDB’s federated query engine lets you write SQL that joins, aggregates, or compares tables from any of your attached sources in one query.
Cross-Engine Joins: Need to compare user data from Postgres with order records in MySQL? Easily combine data from both — DuckDB manages the data movement and query optimization behind the scenes.
Transfer Data Between Databases: Move data from one database type to another (e.g., from SQLite to Postgres) with a single command — powerful for migrations or combined analytics.duckdb.org
Fast Metadata Discovery: DuckDB introspects all attached databases efficiently, giving rapid, comprehensive insights into your schemas, tables, columns, and relationships.
No Export/Import Hassles: Everything happens in memory and on demand — you don’t need to export, import, or duplicate data files.
How DuckDB Handles Federated Queries
Session-Based Attachments: Each Tangent is a DuckDB-coordinated session. The databases you select are “attached” for that session, allowing instant cross-database access.duckdb.org
Automatic Query Orchestration: DuckDB parses your federated queries, figures out the necessary joins and aggregations, and efficiently fetches only the data required from each database.
Native Extensions for More Sources: DuckDB’s pluggable storage and extension system allows connecting with Postgres, MySQL, SQLite, JSON, CSV, and even cloud storage like S3 or HTTPS.duckdb.org
Example (Simplified)
Suppose you have user data in PostgreSQL and transactions in MySQL. Tangents, via DuckDB, lets you run:
SELECT pg_users.name, mysql_orders.amount
FROM postgres_db.public.users pg_users
JOIN mysql_db.sales.orders mysql_orders ON pg_users.id = mysql_orders.user_id
This fetches and joins your data live — no manual data movement required.
Why This Matters
Efficiency: No need for separate ETL jobs or scripting—one query, all your data.
Flexibility: Mix and match database types as needed for analysis and reporting.
Power: Handle analytics scenarios, audits, and cross-checks for complex, multi-source data without leaving SyneHQ.
In Summary
Tangents brings DuckDB’s powerful federated engine directly into SyneHQ’s no-code experience, letting you:
Query, visualize, and analyze data across multiple sources
Explore schemas and relationships effortlessly
Use advanced analytics features without technical setup or maintenance
Whether your data lives in Postgres, MySQL, SQLite, or DuckDB files, Tangents makes it easy to bring everything together for unified insight.
Last updated