Hi, does anyone know how to skip one pipeline when running

kedro run

? I have a model training pipeline which I want to skip since I have a model already working

In

pipeline_registry.py

, you can define whatever set of pipelines you want runnable via

kedro run --pipeline pipeline_name

. You can construct a pipeline object that doesn't have the training pipeline in there.

Yeah but that means I have to add all the non training pipelines by hand?

You can subtract.

pipelines['no_training'] = pipelines['__default__'] - pipelines['training']

Yes, you will need to define what is “training” or “non-training” anyway, there is no way Kedro know it automatically.

Yes but that subtraction is what I needed

You can either separate the pipelines and do the subtraction or addition, or use

tag

@Tomás Rojas how could we have made this more obvious to you in the docs?

I think this non_training = full_pipeline - training is quite handy. It’s more obvious you can

add

pipeline since it’s defined in templates (in the older versions), it may be less obvious that you can subtract them.

I've never used the

tag

option... I'll take a look

💪

Oy maybe a new page where you show how you can make custom pipelines with many pipelines inside

I thought for sure this is documented, but I can't find it anywhere... Maybe it's just in some internal Kedro training presentations? The intermediate training deck had material on this for sure.

Yeah, I haven't read that anywhere either

Cc @Jo Stichbury this could be some inspiration for blog/doc material?