Hey @Bader,
It sounds like you have two questions.
1. When sharing datasets across projects does the dataset stay up to date?
Yes, shared datasets are pointers to the dataset in the original project, not standalone copies.
If you will often have to create new columns in a dataset (in my example screenshot it is a Prepare recipe) then you should 'Share' the output of the recipe. Therefore, whenever you use the recipe to make changes, it will update the schema of the output dataset and all of the projects that the dataset is shared with will receive the new columns/information.
2. Are changes to the original dataset propagated downstream in a project?
Yes, changes you make to a dataset are propagated downstream when you run downstream recipes aka "rebuild the flow". When you click RUN on a recipe or BUILD from a dataset on the Flow you can choose to do so independently or recursively (rebuild some or ALL of the datasets up until a certain point). Screenshot attached.
For more information please check out this documentation about rebuilding datasets or this about the handling of schemas by recipes.
Finally, if you want to automate the rebuild process then you would use Scenarios (learn about them in our Automation course).
Hope that helps,
Emma
