Great question. You can parameterize a batch dataflow to, for example, apply it to another dataset, but you can't edit operations as you would in a dataflow graph.
Batch dataflows are a hardened, operationalizable version of your insights to apply to terabyte-level stores of data. The modifications to create a batch dataflow optimize both the data inputs, outputs and the process in the middle. Changing an operation could impact the entire batch dataflow, risking performance or creating errors.
By comparison, dataflow graphs are a visualization tool that shows the process you modeled to create insights from your data. So, that is where Xcalar added the ability to edit an operation.
You can produce a new batch dataflow from an edited dataflow, as follows:
1) Edit an operation in the dataflow.
2) Run the edited dataflow.
3) Create a new batch dataflow based on the edited dataflow.
Let me know if you have any additional questions about this.