Hello Thunk! I am wondering what (if anything) Thunk just remembers naturally (without Input Properties or anything) from previous steps. Is it a clean slate with every step? Thanks!
Hi Rachel, yes it is a clean slate by design.
The “work item” or “row” represents the memory state for each workflow request.
Each step of the workflow can add information to this memory state. Each step of the workflow can use information from this memory state.
We initially had a default where all properties of the memory state were available as inputs or outputs for each step. The problem with this is that it increases the unpredictability of outcomes. We found that being specific about the input properties and the output properties drastically improves the reliability of outcomes. So we shifted to a default where only some specific properties are inputs to each step and only some specific properties are outputs.
The exception to this is the “Extra Properties” option which is a grab-bag of additional values that vary from one work item to the next. If you enable “Extra Properties” in the input, then you don’t get to pick specific additional properties but rather you get the entire set of them.
Hi Praveen – thanks for your reply. What do you mean by your last sentence (“If you enable ‘Extra Properties’ in the input, then you don’t get to pick specific additional properties but rather you get the entire set of them”)? On my end, “Extra Properties” can be enabled/disabled without affecting what other input properties are selected (which isn’t always all of them).