THE FIELD GUIDE
Make yourself at home.
Start with a small pipeline, bring your own data, then find the guide for the task you want to complete.
Start here
- Installation & accounts — Set up an instance, invite your team and manage access.
- Duckpad in 15 minutes — Build two SQL nodes from scratch, check their data and explain a result.
- Use your own data — Upload a file, check its columns and download a CSV export.
- The Northstar tour — Explore a complete pipeline with CSV, Parquet and reusable components.
- Find your way around — Navigate the canvas, search projects and choose your preferences.
- Troubleshooting — Find help with sign-in, sources, saves, runs and results.
- Dataiku & Alteryx vocabulary — Match Duckpad concepts to familiar Dataiku DSS and Alteryx Designer terms.
Build a pipeline
- Connections — Read and export project files, server folders, S3 and Oracle data.
- SQL syntax — Declare inputs, outputs and typed parameters; write sources and exports.
- Python nodes — Use typed Python nodes, locked environments and managed files alongside SQL.
- Prepare columns & values — Select, rename, cast and clean data with guided forms.
- Reusable components — Group nodes, expose parameters and import reusable flows.
- Project variables — Bind shared values and SQL expressions, with overrides per environment.
- Data partitions — Process ordered batches with bounded memory, retry failed keys and explain their results.
- Document a project — Explain a project, its nodes, output tables and columns.
Inspect and check
- Results, samples & logs — Explore data, save views, compute statistics and inspect the executed SQL.
- Charts — Draw node outputs with Vega-Lite charts declared in charts.json.
- Explain a value — Trace result cells to their source, with explicit coverage limits.
- Data tests — Return failing rows and check data before publication or export.
- Execution messages — Report information, warnings and blocking errors from SQL.
- AI assistant — Ask in context, inspect edits with Undo and confirm runs.
Version and automate
- Version control — Branch, tag, compare, review merges and synchronize your project.
- Jobs — Run reviewed versions and inspect their results and version history.
- CLI & automation — Launch Jobs from schedulers with duckpad run or the HTTP API.
Configure an instance
- Environments — Configure connection values, labels and run confirmations.
- Single sign-on — Connect your identity provider with OpenID Connect.
- Offline extensions — Activate project capabilities and import binaries for offline use.
Reference and development
- Project files & code ownership — Manage project files, imports and the code each node owns or shares.
- Extension import details — Download server binaries, verify imports and troubleshoot loading.
- Architecture — How execution, isolation, persistence and lineage fit together.
- Validation & performance — Measured results, reproducible checks and their scope.
- Scaling benchmarks — Measure 10K–10M rows, PostgreSQL reads and cell-lineage hot paths.
- Contribute to the docs — Edit the website and regenerate its product demonstrations.