my long term goal is to use KNIME on a cluster and be able to run knime projects from the command line. To this end I was trying to create a simple protocol that takes an input filename and an output filename form the command line and reads the inFile and writes its cotnents to the outFile. The important part is that the project does not know how many columns the input file has.
After trying for a while I came up with a solution for sdf files but am struggeling with the csv readers. The basic idea is that the infile and outFile Names are passed via -D java options or environment variables which are then used to create flow variables with the "Java Edit Variable" nodes.
The flow variables are then used to define the input and output filenames in the read and write nodes.
However the "File Reader" and "Variable Based File Reader" both read in the number and type of columns during the configuration of the node and give an error when applied to files with more or different columns.
Did I miss anything or is this simply not possible with the current nodes? It seems such an easy task.
I have attached my example project.
Thansk for any input,