I’m not sure if I can use knime for this project I am implementing with a recursive loop. It starts off fast and then progressively slows down after the 10th iteration. Eventually, I run out of heap space even though I maximized it in the .ini file. Can someone please advise?
it would be more helpful if you described the dataset, the task (why recursion?), and possibly uploaded the workflow with the data/example data.
I’m running a 48 hour ahead forecast algorithm. I want to do recursion so I can feed the predictions for the n-1 hour into the n hour features to capture the time correlations in and improve accuracy of predictions. I am looping over a table with 48 rows and 230 columns. It doesn’t seem like memory should be an issue but it definitely is. I’ll see about uploading an example with data.
I have an example. Can you please tell me how to send it to you?
Just a quick update: The problem appears to be with the joiner nodes. When I take them out, the loop runs well. I do need them in the loop, though, unfortunately…
Problem solved! I used the column appender node instead of the joiner node.
We also have a new joiner node, which is a lot faster. You can find it here Joiner (Labs) Demo
For workflow upload we are recommending the hub.knime.com There you can share your workflow and others can see and access them.
you could upload workflows to this thread, in the future.
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.
This is probably due to long RowIDs generated due to Joiner in recursive loop.
See here for more: Recursive loop performance issue