Improve the column filtering feature to use custom IColumnFilter implementations

For now, APE implements column filtering based on column names specified in the excludeColumns attribute of the @ShouldMatchDataSet annotation (introduced in ARQ-772). This could be extended to allow custom IColumnFilter implementations to be plugged in during the verification phase.

The use of IColumnFilter implementations is described here: http://www.dbunit.org/faq.html#columnfilter