Text mining: How to remove vector words from the output

fines

Hi all:

after applying a learned model to an unlearned data set, my objective is to store the original atributes plus the predicted clasification to an Excel output file.

The problem comes when I connect the lab port of the Apply Model operator to the imp port of the Write Excel operator,  becouse the model creates also an atribute in form of Excel column per each of the words on the words vector, so at the end the Excel file contains too many columns to be writen.

Somebody knows how to reduce the columns in the output excel file to just the original attributes columns, the prediction column, and the confidence columns  and remove the attributes that represent the word vector info?.

Thanks in advance