Overview
Review the execution contracts and interface-specific input formats you need when writing Python or SQL in D.Hub.
Choose a reference
| Page | What it covers |
|---|---|
| Python code reference | The pipeline Python code node's run function, inputs, options, shared values, outputs, available packages, and examples |
| SQL reference | Differences between the input tables used by pipeline SQL nodes, dataset exploration, and dashboard SQL mode, plus ClickHouse functions |
Distinguish the interfaces
| Where you write code | Reference | Input convention |
|---|---|---|
| Pipeline Python code node | Python code reference | A Polars DataFrame passed to run(input, *, options=None, contexts=None) |
| Pipeline SQL node | SQL reference | The input connection alias, or input by default |
| Dataset Data tab | SQL reference | The current dataset's physical table name |
| Dashboard widget SQL mode | SQL reference | The selected data source's database.table |
Python and SQL do not share the same input and output contract. In particular, the input table name applies only to pipeline SQL nodes.
Related tasks
- See Add nodes to place code in a pipeline.
- See Collection code to create code.