You could try some different export formats besides the original KNIME table format. Excel I think is quite slow because it has to be specifically processes. If you want to have better performance you might have to use some streaming or a different format. Most of the formats I compared would preserve your variable formats.
R is there as a test if this might handle your memory issues in a different way. To speed things up obviously it would not make sense to first send the data to R and the export it.
aside: yes I am aware a true comparison would need the machine to run the exact same tasks while doing this. So it is more an estimation but maybe still useful.
Expanded @ipazin workflow:
2019_05_14_Writing_Excel_File.knwf (409.2 KB)