How can I configure an n8n workflow to continue execution when only one branch outputs data?

When using the Merge node in “Combine by Position” mode, n8n requires that each input provides a matching item so that the merge can occur. This means that if one branch is delayed or produces no items, the node won’t output until it sees a corresponding item from the other branch. There isn’t a built‐in configuration that tells the Merge node to “fire” as soon as one branch has data.

Some approaches to work around this limitation are:

• Use a different merge mode:
  For example, the “Append” mode does not require inputs to exist on all branches; it concatenates the items from all inputs. (Keep in mind that this mode will not maintain the positional order between the branches.)

• Restructure the workflow:
  If you need the subsequent nodes to proceed as soon as one branch produces output, you can avoid merging the branches too early. One strategy is to let each branch run in parallel until you need them to interact. Later on you could merge the data (using a mode like “Choose Branch” if only one branch’s data is needed, or some other method) or use additional logic (for example, an If node or a custom Code node) to combine the outputs.

• Custom handling for uneven items:
  In cases where you’re merging data streams with an uneven number of items, you might consider preprocessing the branches to ensure that the “slower” branch always produces a placeholder item. Although you mentioned that setting the Fill Markdown node to always output data isn’t ideal for more complex branches, a similar idea (implemented with conditionals or a Code node) might give you more granular control over when each branch outputs its data.

In summary, there isn’t a direct configuration option in the Merge node (when using “Combine by Position”) to allow early execution when only one branch produces output. You’ll need to either choose a different merge strategy or adjust your workflow design so that you don’t depend on the Merge node waiting for data from both branches before continuing.