Difference Meta Node and Sub Node

Hi folks!

Being absolutely new to Knime, I came across the sub node / meta node functionalities to structure the dataflow code. Actually, I couldn't derive from the available documentation what the differences between these two types of subroutines are. Can anybody help?

What I found so far is located here: http://tech.knime.org/wiki/meta-nodes

Thanks in advance!

- Darak


Hi Darak, 

The main difference now is in the flow variable scope.  For example, you can hide certain flow variables in a subnode and they won't propagate down stream.  More interesting features and better documentation will be coming with KNIME 2.11

Hi Aaron,

I has a look at KNIME 2.11 but found no further documentation. Can you give me a hint where to find it? E.g. I couldn't figure out how to create a sub node.



As far as I understand it, a subnode acts pretty much like a node - you can either execute it or reset it.  A meta node is more like a workflow - you can execute bits of it, all of it,none of it etc., but it is a convenient way of organising large workflows or repeatedly used sets of nodes.

There may be other differences too!


I could not find any documentation on subnodes either. Where can we find it?

The difference between metanodes and subnodes is confusing, why have two different types, anyway?.

It starts with the UX inconsistencies: to open a subnode you need to use the context-menu (right-click vs double-click), the behaviour @runtime is different, in/out-put ports are represented in another way.

It looks like the raison d'être of the subnode is to shield off a part of your workflow. It's like a concrete wall for workflow-variables: nothing comes in, nothing gets out, not even if you'd want to (by using variable ports).