Daniel Thom

@andrewrosemberg I understand that you might be interested in this feature. Please review and add comments if you'd like. I can try to accommodate your needs.

This proposal is only useful if you need to swap out time series data for each scenario. If you only need to change a few components there are better things...

@andrewrosemberg Have you considered calling `remove_time_series!` and `add_time_series!` on each of the three generators for each scenario. You could maintain one system throughout. If you were to do something like...

The component is required because the user may be using the `scaling_factor_multiplier` concept. In that case the time series instance stores a reference to a component method that will be...

No, that's what you would have to do. It's even more cumbersome for subtypes of `Forecast` because the data is actually stored in a `SortedDict`. The user would have to...

This is certainly not a good idea unless the user understands the internals. If the generators are not attached to a system then they need to change the name and...

This means that the descriptor file does not contain the struct being used. We have an oversight in this process. A user might want validation of a struct without doing...

My earlier statement was incorrect. You can define a validation descriptor for a struct that is not auto-generated by adding an entry here: https://github.com/NREL-SIIP/PowerSystems.jl/blob/master/src/descriptors/power_system_structs.json#L9581. There are a few options: 1....

Yes, the issue no longer occurs.

@felipenoris I traced through the code and understand why this issue occurs. It's tied to `Base.iterate(itr::SheetRowStreamIterator, state::Union{Nothing, SheetRowStreamIteratorState}=nothing)`. The first time the function is called it opens a file handle...