H2O Error Job Crashed Unexpected - "No rows left in the dataset after filtering out" NA

Hello! I have encountered the following error, and am not sure as to the cause of this crash:

ERROR H2O Generalized Linear Model Learner (Regression) 5:16460:0:16453:351:0:345:16464:17 Execute failed: Job crashed unexpected. Cause: org.knime.ext.h2o.exception.H2OJobCrashedException: Job crashed unexpected. Cause: No rows left in the dataset after filtering out rows with missing values. Ignore columns with many NAs or impute your missing values prior to calling glm. See log for details.

The error log was blank when I checked. This node is nested in a Feature Selection (genetic algorithm) loop, and I can clearly see that prior to the GLM Learner, my data frame contains no missing values.

I have two workflows executing simultaneously, in different KNIME workspaces, and am not sure if this error may be an issue with memory. For the time being, I have configured error handling nodes in my workflow, within the genetic algorithm loop, to execute a branch that does not include the GLM Learner. This alternative branch appears to execute just fine, so I am not sure what is the cause of this crash.

I am using KNIME Version 4.6.3, I am running this workflow on a pc (Microsoft windows) with the following specs:

  • 64 GB of RAM (152 GB virtual memory)
  • Intel Core i9 -10900KF CPU @ 3.70GHz

My data frame has just over 50,000 observations and slightly over 240 columns / variables. Below is a screenshot my workflow:

(Note: the workflow is currently executing, however after some time I see the error for the top branch)

Hi According to the message that is due to the fact that one column in your dataset contains missing values but if you already checked that then you might want to have a look at the h2o doc for the GLM.
Maybe that could enlighten here.

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.