conditional and unconditional margins after two-part model

Hi Statalit
I use Stata 14.

I'm trying to run a two-part model (first-part : logit and second: glm/poisson/log) and the perfectly works.
To show my results, I would like to show margins of the sample of the second part (i.e. the conditional margins) and then the unconditional margins.
When I use I assume it gives unconditional margins for the whole population.
But It seems I cannot use because the value stored is the sample of the first part.

I guess I could use a step by step approach to independently run my GLM model in patients with non zero values and then use margins.

But, does another easier mean exist?
Thank you very much for you help