I had a deep look at your uploaded workflow and unfortunately it still seems to be corrupted although partially working.
For instance, I realised that some metanodes work but internally, they are corrupted (even if still working!) with lacking connexions, as shown here below:
I guess that the -Group Loop Start- node is as far as the workflow can work in this state given that it is corrupted but definitely it cannot run beyond this start loop node. Maybe it is not even a problem with the -Group Loop- node but with all the corruptions before in the workflow.
I do not think there is a solution other than creating a fresh workflow with all the nodes and metanodes that are not corrupted and work, and replace by new nodes and metanodes those that seem to be corrupted.
Sorry this is as far as I can go. Please reach out if you have any questions though. I will be happy to answer them if I can.
Ok, I can understand.
The Metanode (in your last screen) is working because the node you see is the backup node on the left side of the metanode. if you go at the beginning of the metanode you can see the nodes that are working. I donât know why but every time I open this metanode it makes me see this part that is not the core node.
This node alone on its own was confusing and led me to a misinterpretation of this metanode which eventually seems to work. I understand now and I prefer to withdraw my last post to avoid misleading future readers.
I will have a further look to your workflow in case I find anything else.
@pnotova from what I see the âwireingâ of the internal loops is off. Where you marked it as not working there is a âColumn List Loop Startâ that end nowhere and will not work I fear. You will have to complete the logic of the loops. And I would also recommend to get rid of the âadditionalâ nodes that are stored on the sidelines. It might be that they are also being tried to execute and cannot work.
Thanks @mlauber71 ! Indeed this is the problem. I just checked and there were plenty of unconnected nodes and this -Column List Loop Start- node was left without Loop End node connected which is redibitory in the case of KNIME.
I cleaned up all the unconnected nodes in your workflow and now it doesnât hold in a queued state but there are new errors that appear in your workflow, for instance in the following metanode:
@pnotova Maybe you can take it now from here, run it and check now why it is failing because of not finding the expected columns in the -Column Aggregator- node.
I will try to clean the flow you are right it is kind a mess and I will let you know as soon as possible!!
@aworker yes I think that the flow gives you errors because of less data. I didnât give you the whole data set because it is huge and it was not necessary for our scope!
@pnotova you will have to make sure all the loops are correctly wired. You can select one loopâs current focus with right click. Sometimes it can make sense to put a Cache node in fron of the loop. Also there have been cases where it was best to encapsulate the inner content of a loop in a component in order to make sure what to execute when if you have nested loops.
On performance and KNIME in general I can offer this article:
Thankâs again.Itâs working now!!! I will keep my flows cleaner:slight_smile:)
Promise!
Thank you again and see you somewhere in the Knime community!
