Variant subystems with mixed discrete/continuous?

I am trying to use a variant subsystem to switch between a continuous block and a discrete block - a derivative block and a difference block, respectively. When the variant is set to the derivative block, the diagram will be used to run a simulation directly. When the variant is set to the difference block, the diagram will be used to generate code.

In principle, this should work, since only one variant is active at a time. When I generate code, I want the continuous variant to be left out completely, as it will be inactive. However, when I try to generate code, it complains that there are continuous states, so it can't (presumably because the derivative block is still there, even though it is inactive.) Is there some way that I can configure it to completely ignore non-active blocks, depending on how my variants are set?

[Azzi Abdelmalek]

what do you mean to generate code?

[Jason]

I am referring to the Simulink coder... ctrl+B

Answer 1

[TAB]

See Code Generation of Variants.

I think enabling Override variant conditions should work for you.

[Jason]

That option, as I understand it, merely removes the Variant Conditions from the picture, and allows you to specify a variant directly. My Variant Conditions are operating properly, and it shows the correct one as "true" in the variant subsystem parameters, so there's no reason to use this option.

Answer 2

[Guy Rouleau]

I would be surprised if this is because of the derivative block. Usually, inactive variant should completely disappear.

- Ensure the option "Generate preprocessor conditional" is disabled. Otherwise both variants will be present... this is the only reason I know for 2 variants to be present.

- Use sample time colors or Simulink.BlockDiagram.getInitialState to identify which block has continuous state in your model.

[Jason]

I have the "generate preprocessor conditionals" disabled, but it appears that when code is generated is still cares about the inactive variant. I don't know why.

Can you elaborate on the sample time colors? I don't understand what you are saying here.