Thanks, that is much faster.  I don't quite understand it, but that's OK
for now.

Thanks again,

Matthew



On Sat, Feb 25, 2012 at 9:00 PM, Matthew Dowle <[email protected]>wrote:

> Hi. There have been thoughts of adding .I for that. In the meantime
> something like this should be faster:
> DT[DT[,X1==max(X1),by='x,y,z'][[4]]]
>
> > I  know that we shouldn't subset .SD, but I would like to return the row
> > from each group that corresponds to a max value in one of the columns.
> >  Here's how I'm doing it currently:
> >
> > maxrows<- df.dt[, .SD[which(X1== max(X1)),], by = list(Var1, Var2, Var3)]
> >
> > Is there a better way?
> >
> > Thank you kindly,
> >
> > Matthew
> > _______________________________________________
> > datatable-help mailing list
> > [email protected]
> >
> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/datatable-help
>
>
>
_______________________________________________
datatable-help mailing list
[email protected]
https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/datatable-help

Reply via email to