Hi 🙂 No, that's not possible (in a similar way to git). Can you please tell me what is your use case regarding that?

i am still in the checking phase. we have a long process that has sub processes in it. if we have a failure at the end it would be nice to be able to commit some of the smaller processes. for example we are using dbt. we have a dag with 5 models (each with tests), if the final test fails we might want to merge some of models into master, since they passed

But note that there is no "current" branch. So perhaps this might help: Look at other branches using their pathnames. You can tag as you go or just use "ref expressions" to access non-tip revisions, e.g. the expression

process-branch~2

is exactly the grandparent commit of whatever

process-branch

branch is now looking at.

not sure that i understand. if i tag different branches, can i bring in all the information to one place?

You should be able to merge from any ref, not only a branch.

I believe that with dbt the flow might look a little different. In a naive way we could have done a commit after every model execution, and then merge only the models that passed. However, dbt executes the models in parallel (as long as they are independent), and writes to one fixed schema. This is the reason we can't commit after each model execution. Therefore, one way to commit only the passed models is to delete the irrelevant models (those who didn't pass), and then commit all passed models together. Another possible option (which we didn't explore yet) is to use dbt commands in order to manually manage the flow.