Coming from…
Coming from Flowfile
Flowfile and Daedalus are close cousins. Both are visual ETL apps where you wire nodes on a
canvas, and both let you take the result out as Python code. Most of what you do in
Flowfile has a node with a similar name here.
What feels familiar
- Nodes on a canvas. Read, transform and write steps connect the same way. Select a node and set it up in the Attributes panel on the right.
- Results under the canvas. The selected node's output shows in the Data tab, with Map, Chart and Metadata tabs next to it.
- Code when you need it. A DuckDB Query node takes SQL. A Code Gen node writes Python from your description. The pencil on any node opens its Python for editing.
- Export to Python. Export as Python library turns the whole workflow into a Python package you can install and run without the app.
What's different
- The engine is DuckDB. Most steps run as DuckDB SQL. A few steps that have no SQL form, such as Exponential Weighted Mean or Write Excel, use Polars.
- Data comes in through drones. Every source is listed in the Drones panel on the left. A drone is a small helper that reads the file, folder or database for you; drag a source from the panel onto the canvas to use it.
- Geospatial is built in. Daedalus reads and writes geo files, joins by location, routes on roads and draws maps without plugins.
- Reports and deliveries. Workflows can end in a PDF or PowerPoint report, an email or a Slack message, not only a file or a table.
No equivalent yet
Some Flowfile tools have no Daedalus node yet. Searching for them shows a hint instead of
an empty list. These are the ones you are most likely to miss:
| Flowfile name | What to do instead |
|---|
| Apply ML Model | No model scoring; Linear Regression can output predictions itself |
| Evaluate Model | No metrics node; compute errors with Add Column + Aggregate |
| Kafka Source | No streaming sources; land the topic to Parquet and read that |
| Explode hierarchy | No BOM/hierarchy explode; write a recursive CTE in DuckDB Query |
| Google Analytics | No GA4 connector; export the report and read the CSV |
Tips
- Search with Flowfile names. Group by, Unique and Text to rows find Aggregate, Distinct and Split to Rows.
- Polars code becomes SQL or Python. If your Polars expression is mostly column maths and filters, a DuckDB Query node is usually the shortest rewrite. For anything else, describe it to a Code Gen node.
- Group steps you reuse. Select a few nodes and wrap them in a named function: a frame you can collapse into a single node and run from there.
- Keep column formats between projects. Write Daedalus Table saves a table with its column formats, and Daedalus Table Reader reads it back in another project.