MotherDuck

Instant SQL is HERE 💥

NEW DROP: MotherDuck INSTANT SQL 🔥 Speedrun ad-hoc queries as you type. Hamilton Ulmer

Hamilton Ulmer mic drops INSTANT SQL at Data Council 🔥🔥🔥 Bend reality with real-time SQL results... Speedrun ad-hoc SQL queries as you type... Type, see, tweak, repeat... Get SQL superpowers - it's 2025. Learn More: https://lnkd.in/dhpYcWgx

We chase down pipeline failures, debug Spark jobs, and monitor orchestration — all through log files. But here’s the twist: Logs are data. In our latest post, Simon Späti breaks down how you can stop grepping and start querying logs using SQL with DuckDB and MotherDuck — including: - How to parse messy JSON and syslog files like structured tables - When to materialize vs. query in place - Analyzing millions of rows without leaving your laptop - And why log analysis is a competitive advantage for data teams 👉 Read the full guide: https://lnkd.in/eP3yxC2m #duckdb #dataengineering

Scaling DuckDB is at the heart of what we do at MotherDuck. Sometimes it's necessary to power a customer's business. Sometimes it's just important to help with the 'noisy neighbor' problem. Sometimes it's necessary to mitigate scale anxiety. MotherDuck co-founder Ryan Boyd 🐤 wrote a post on how we scale-- vertically, horizontally, through per-user tenancy and with Dual Execution. Most importantly, we introduce the term we use for DuckDB instances: Ducklings! https://lnkd.in/g4bTwaEB

Data engineers know the frustration: building pipelines often involves slow feedback loops due to complex tools and data dependencies. Unlike the quick refresh cycles in web dev, we're often stuck waiting. But what if AI could directly interact with our data sources during development? Enter the Model Context Protocol (MCP) – an emerging open standard connecting AI copilots (like Cursor, GitHub Copilot) directly to tools like databases (hello, DuckDB!). In our latest blog post, we explore how MCP can significantly shorten development cycles by allowing AI to: ✅ Query & inspect data sources directly (e.g., run DESCRIBE on Parquet files). ✅ Understand schemas BEFORE generating code. ✅ Test transformations iteratively. Mehdi Ouazza walks through a practical example using MCP with DuckDB, Cursor and dbt to build pipelines faster and more accurately. Stop the waiting game and accelerate your data workflows. https://lnkd.in/ecJBSsG6 #duckdb #dataengineering #mcp

Looking to leverage AI in your data workflows? Yesterday, we had a series of lightning talks with practical tips on how to use AI. 📒 Enrich your data catalog with AI-powered workflows (Evelyn Wou) 🔄 Shorten your data engineering test-deploy cycles using MCP feedback loops (Mehdi Ouazza) 📊 Leverage Cursor in development of BI charting (Archie Sarre Wood) 🧹 Use LLMs inside of SQL for cleaning up messy sales data (Nate Sooter) 🎬 https://lnkd.in/gZv9tu58

🚨 New Month, New DuckDB ecosystem Roundup! April’s edition is packed — here’s a sneak peek: 👤 Featured Community Member: Shoutout to Caleb Fahlgren from Hugging Face! He shows how the new DuckDB Local UI makes it effortless to explore hf:// datasets right from your laptop. Oh, and did you know DuckDB also powers dataset discovery behind the scenes at HF? 👀 🔥 Top links this month include: • DuckDB + Streaming with SQLFlow • Hybrid vector search with Quiver (DuckDB + HNSW + Arrow) • Speeding up S3 queries with local cache • DuckDB UI tips, from notebooks to column explorers • Security + startup time improvements with static compilation • A Yazi plugin to preview data files in your terminal • SQL tricks like renaming fields in SELECT * across JOINs • And a guide for Pandas/Polars users mastering DuckDB SQL 💛 Big thanks to the contributors and DuckDB community! https://lnkd.in/eRZeeeAh #DuckDB #DataEngineering

Forget the hype—come see how folks across the data world are putting AI to work in their everyday jobs. ⚡ We're hosting 3 lightning talks plus a Q&A on how real data people are actually using AI We're bringing together: 🏎️ Mehdi Ouazza on Speeding up the the data pipeline test-deploy cycle with MCP feedback loops 📈 Archie Sarre Wood on configuring Cursor to maximize productivity in data app development 🧹 Nate Sooter on Cleaning up messy CRM data using LLMs right in SQL 📅 Wednesday, April 9 at 9:30am PDT 🔗 https://lnkd.in/gQEmt63z

DuckDB recently shipped a new local UI. It's super powerful and works with any of the 384k+ datasets on Hugging Face. Here's a post on how to use it! https://lnkd.in/g34kfwyw