jan
May 7, 2023, 6:37am
2
That is not a bug but rather currently expected behaviour. It is however planned to change that with the upcoming n8n version 1.0. There is already a PR:
n8n-io:master
← n8n-io:workflow-version-2
opened 08:43PM - 29 Sep 22 UTC
This PR contains two changes that make it possible to configure how nodes with m… ultiple inputs behave. It is now possible to control if incoming nodes get forced to execute (currently default), and it is possible to define which inputs are required.
Important to mention that the setting `forceInputNodeExecution` always force executes all nodes from all inputs rather than just the ones defined in `requiredInputs`. (Background: this specific behavior is currently not required and may never be, so no need to spend time on it and unnecessarily complicate the code)
## Example of how new versions of the Merge-Node will behave now
Current data processing behavior:
![Screenshot from 2022-10-03 22-34-26](https://user-images.githubusercontent.com/6249596/193677656-583617d9-e294-45bd-b078-1a2fde0f826d.png)
New data processing behavior:
![Screenshot from 2022-10-03 22-35-42](https://user-images.githubusercontent.com/6249596/193677668-f21312a1-3dc3-47a0-b858-5cda1b9f4843.png)
Community topics:
https://community.n8n.io/t/merge-data-node-getting-triggered-even-without-data-input-from-both-source/22174
https://community.n8n.io/t/getting-error-in-if-node/22179
https://community.n8n.io/t/if-node-is-true-and-false-in-one-execution/3169/5
https://community.n8n.io/t/if-node-continues-to-both-branches-even-though-it-is-false-and-data-is-on-the-false-side/21908
https://community.n8n.io/t/merge-wait-seems-broken/18725
https://community.n8n.io/t/unexpected-behavior-from-unvisited-nodes-in-a-switch-and-merge-workflow/10278
https://community.n8n.io/t/merge-node-suggestion-add-option-to-disable-triggering-of-previous-nodes/9931/13
https://community.n8n.io/t/workflow-nodes-execution-order/13546/6
https://community.n8n.io/t/bug-merge-node-makes-all-preceding-nodes-execute/25939