Hi, I don't mind as I don't use transform() or within(). At least, I don't use them directly. Packages I pass a data.table to might I suppose. Sounds like a good idea to me though. Matthew
On Thu, 2011-03-31 at 12:29 -0400, Steve Lianoglou wrote: > Howdy, > > Would it make sense to change `transform.data.table` to ensure that > the key columns are maintained after the `transform`-ation? > > My initial thoughts is that this is "fine" (ie. won't break anything) > as long as none of the columns transformed are elements in the key -- > if they are, I guess it's not clear what to do. One option is to drop > the keys at this point, another is to add a `.force.keys=TRUE/FALSE` > argument after the dots, eg: > > transform.data.table <- function(`_data`, ..., .force.keys=TRUE) > > which would reset the key columns to what they were, even if it > resorts the data.table. > > Thoughts? > > -steve > > btw - and I guess whatever makes sense to do here would also make > sense to do in the `within.data.table` function as well. > > _______________________________________________ datatable-help mailing list [email protected] https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/datatable-help
