Error using Logistic Regression (Learner) with 65.000 rows

Hi,
I'm executing a basic predictive model with Knime and I'm using "Logistic Regression (learner)" node and I haven't any problem when I use between 100 and 1000 rows, but when I use files with 65.000 rows Knime show me this Error:
 
===== Executing workflow Workflow_Test_3.zip =====
WARN     CSV Writer      Selected output file exists and will be overwritten!
WARN     File Reader     Node created an empty data table.
ERROR    Logistic Regression (Learner)   Execute failed: ("NegativeArraySizeExce
ption"): null
ERROR    KNIME-Worker-1 Logistic Regression (Learner)    Execute failed: ("Negat
iveArraySizeException"): null
WARN     CSV Writer      Selected output file exists and will be overwritten!
Finished in 4 secs (4126ms)
========= Workflow did not execute sucessfully ============

Why this happens, when increase the number of records to analyze?

Thank you.

Hi,
 
I tried execute a job very much simple that the last, but use the same file(65000) Lines(Rows).
 
The workflow must read a text file and order all rows based a column and write a file with this results. (See the image attached)
 
The execution Log is:
 
!ENTRY org.eclipse.core.resources 2 10035 2013-07-25 12:11:44.189
!MESSAGE The workspace exited with unsaved changes in the previous session; refr
eshing workspace to recover changes.
WARN     KNIME-Worker-0 File Reader      Node created an empty data table.
WARN     KNIME-Worker-0 CSV Writer       Selected output file exists and will be
 overwritten!
WARN     KNIME-Worker-0 Table Writer     Selected output file exists and will be
 overwritten!
WARN     KNIME-Worker-1 Sorter   Node created an empty data table.
WARN     KNIME-Worker-1 CSV Writer       Selected output file exists and will be
 overwritten!
WARN     KNIME-Worker-1 Table Writer     Selected output file exists and will be
 overwritten!
WARN     KNIME-Worker-2 Table Writer     Selected output file exists and will be
 overwritten!
INFO     main BatchExecutor      Workflow execution done Finished in 4 secs (476
6ms)
INFO     main BatchExecutor      ============= Workflow executed sucessfully ===
============
 
Also I exported the workflow for reuse with different files (which have the same format and the same file name), for this reason I run in command line mode (batch mode), using the following command line:
 
knime.exe -noexit -consoleLog -nosave -application org.knime.product.KNIME_BATCH_APPLICATION -nosplash -workflowFile="C:\temp\prueba1\job\Test1.zip"
 
Finally, this workflow works well with few lines files (100 or 1000 lines), but with a file of 65000 lines appear this log line "WARN     KNIME-Worker-0 File Reader      Node created an empty data table." and the result file is ever empty.

 

Is this the correct forma of use Knime?

 

Thank you.