Convert BCUT matrix to columns

KNIME Analytics Platform
1

Hi,

Among the CDK Molecular Properties are BCUT descriptors, which are compiled in a multicomponent vector, like so:

[14.003078088666154, 12.15008154167782, 12.004914369656012, 11.99900000000001, 11.999, 11.999, 11.998999999999999, 11.998999999999995, 11.849999999999998, 0.08851914963640023, 0.027519749784018485, -0.030796102675157283, -0.045605651031546235, -0.05612139135851023, -0.058096780534798095, -0.061847326533073864, -0.21159626793566777, -0.3238090794218826, 9.281630202260658, 8.49519464240508, 8.0670056460103, 8.065000000000001, 7.513343100992997, 7.452157390691937, 7.084145529173876, 6.182343125655294, 5.905086612809867]

I would like to be able to convert this vector into separate columns, in order to be able to perform PCA on them, but don't know how to do the conversion. I've tried the cell splitter but it doesn't recognize to BCUT column type.

Any suggestions highly appreciated!

2

Hi, I think it is possible via string transformation in any scripting language node. For example in Ruby Snippet it looks like:

row[0].to_s.delete('[]').split(',').map(&:to_f).
  reduce(Cells.new){|row, item| row.double(item)}

The specification of an output table must contain an appropriate number of columns.

3

the simple way is to use split collection column

 

fab

4

Many thanks, this seem to do the trick!

5