I created this workflow all_columns_to_one.knwf (21.6 KB) it uses a Column List Loop to process every column one by one . To have them into one column a the end, all columns are renamed to a single name. And in case of missing values (?) they are filtered. Hope this helps.
you can use the unpivot node. After unpivoting you get 2 Columns. One column contains the column headers of the previous table and the other column the corresponding values. The only thing what you additional need is at least one unique identifier for your table rows in your source table e.g. ROWID.