Currently headers are added to KnoxShellTable with any trailing or leading spaces intact. This makes any operations on the table that take a header name as a parameter difficult to use. There is no way to determine whether there are such spaces in the table even when rendered due to the padding in the cell.
I can't think of any reason that such spaces need to be retained even for CSV files when they exist in the source file.
This patch will trim the spaces so that header name based operations are more easily done and coded.