Doesn't that exist more or less?

https://pypsa-eur.readthedocs.io/en/latest/configuration.html#adjustments

I think like this it is more flexible, because you can vary everything (e.g. also component loador other attributes than costs and potentials). Also it allows to modify depending on the investment period.

But one should decide for either one of the definitions and also clean up the settings in the config before merging and move them to adjustments or vary...

In my humble opinion I think it would a little confusing to have both adjustments and vary config sections - arguably there is already an intimidating number of different ways of adjusting the configuration.

Thanks a lot for your suggestions @koen-vg and @fneum . I still think it is useful to be able to more flexibly modify the components and attributes. For example, with the current implementation in the master branch the load cannot be modified and only specific attributes. Also it is not possible to have a modification depending on the investment period, which can be very useful. I have adapted now the existing function maybe_adjust_costs_and_potentials to work more flexibly, including the option to vary by a factor or to set an absolute value. Let me know, what you think about this!

I definitely see the appeal of being able to adjust any attribute, also depending on the investment horizon (and choosing relative or absolute adjustments)! I think it made a lot of sense to build this into the adjustments config.

If you ask me, I would at this point just get rid of the maybe_adjust_costs_and_potentials function (calling modify_attribute directly), since it really only serves as thin scaffolding to modify_attribute. Can just be less confusing to people that don't know the history of the maybe_adjust_costs_and_potentials function. But even without this change the PR looks good to me :)