On Wed, Feb 08, 2006 at 05:27:55PM +0800, John Darrington wrote: > I'm wondering if there's a better way we can implement the /MISSING > subcommand which appears in nearly all of the stats commands. > Currently, it's up to each procedure to explicitly implement it, which > is both a pain and error prone. > > Perhaps we could pass a parameter to the case_reader when its > constructed so that it returns only the appropriate cases?
This is a good idea. Sometimes procedures need the missing values, but usually they are dropped. Passing a parameter to the case_reader would be an easy way to take care of both situations. -Jason _______________________________________________ pspp-dev mailing list [email protected] http://lists.gnu.org/mailman/listinfo/pspp-dev
