Hey,
I have a workflow with a R Snippet (Interactive R Statistics Integration) which works in KNIME 3.2.1, but not in KNIME 3.3.0/1. The error message in the console is the following:
Execute failed: R evaluation failed.
When executing the script within the node's dialog, the error message is more detailed:
ERROR: R evaluation failed.ERROR: R evaluation failed.ERROR: eval failed, request status: evaluation successful, but returned object is too big to transport
The incoming table has 4.6 million rows and 7 columns.
The R code used does not play a role here, because knime.out <- knime.in is not working either.
Please help!
Best,
Marc
Hi Marc,
Hiks, those types of posts always cause some stress in the group ;-)
I just did a code diff between 3.2 and 3.3 and the only two changes are the bump in the version number and the change required to address:
AP-6713: Rserve receiving buffer size limit should accept 0 as input
KNIME now also sets a 'maxsendbuf' on RServe (and permits 0 values in the R preferences). I doubt that this is the origin for the problem you are seeing. It might be the data is just large and it always was some kind of a corner case?
Can you double-check the value in the KNIME-R preferences? (0 should set it unlimited... )
- Bernd
Hey Bernd,
the number was the same in both versions (256 MB). When changing it to 0 in 3.3.0/1 it works (again)! :-)
Thanks,
Marc