> For the complete documentation index, see [llms.txt](https://docs.dataspace.ch/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.dataspace.ch/platform/transformation/code.md).

# Code

The code repository can have two types of code files: regular Python scripts and transformations. For a transformation to be identified as such, a function called `transform` returning a LazyFrame (or DataFrame) needs to be present. The function will be called automatically when appropriate.

```python
import polars as pl

def transform(source_1, source_2):
    return source_1.join(source_2, on="id")
```

The `transform` function can have multiple parameters. A parameter to the `transform` function needs to have the same name as one of the transforms. By declaring a transform as a parameter, the dependency graph will be updated automatically, the transform will be read, and the DataFrame will be passed in as a parameter to the `transform` function.

The following example shows how a join transform declares a dependency on `source_1` and `source_2` by specifying those transform names as parameters for the `transform` function.

<figure><img src="/files/gXmGTFMmuFWXpme7LAUX" alt=""><figcaption></figcaption></figure>