Modifying a Dataset by Recipe in Driverless AI
- Predictive AI
- H2O Driverless AI
This lesson walks through the Modify by Recipe button on the Datasets page in Driverless AI, which offers a data recipe loaded from a URL, uploaded directly, reused from one already applied, or written live in the editor. It shows the Live Code option, where an example Python script keeps only the active, uncommented lines and can be edited to filter or transform the dataset in any way needed. The lesson finishes by applying the recipe, which creates a new dataset row with the reduced column count and confirms the transformation worked as written.
Key takeaways
- How the Modify by Recipe button offers four ways to apply a data transformation to a dataset
- What the Live Code editor does with commented versus active lines when a recipe is applied
- How applying a recipe creates a new dataset row that reflects the transformation just written
- Why writing a custom Python recipe gives more control than the platform's default dataset options
Start owning your AI