Kedro is an open-sourced Python framework for creating maintainable and modular data science code.

Kedro

Hi, I am able to pass parameters to my Kedros pipeline at run time via the --params CLI as,

*kedro run --pipeline=experiment_one --params test_fold: 4*

Here the same pipeline can be run with different test_folds.

The challenge is how to use this dynamic variable "test_fold" to name output files on the fly?

I am looking for a solution something like,

node(func=test_function, inputs=[..], outputs=*'results_%s.parquet' % (params:test_fold)*, name='test_function')

Assume my catalog.yml has the results_*.parquet files already pre-defined.

Is this possible?

(Disclaimer:I think this is not a good idea to do this, you should use an experiment tracker like mlflow, W&B, Neptune ... to store the different results of your experiments)

That said you can use the `runtime_params` resolver to achieve what you want : 

`# catalog.yml`
`my_output:
    ...
    filepath: results_${runtime_params:test_fold}.parquet`

See: <https://docs.kedro.org/en/stable/configuration/advanced_configuration.html#how-to-override-configuration-with-runtime-parameters-with-the-omegaconfigloader|https://docs.kedro.org/en/stable/configuration/advanced_configuration.html#how-to-override-configuration-with-runtime-parameters-with-the-omegaconfigloader>

<@U0482AVABUN> Thank you for your reply. Let me experiment with this option.