Not pdiff because i) psum(x,-y,na.rm=TRUE) would do that and ii) diff is
quite unlike -. Yes, pprod too, but not pdiv (or precip) because
pprod(x,y^-1,na.rm=TRUE) would dominate that.
what about equivalent functions for ^, %%, %/%, &, and | ?
I like the suggestion, but not as useful as psum and pprod. It would
probably be going too far to add those too. Plus in ?groupGeneric, under
section 3, there are 7 functions listed :
min, max, sum, prod, range, all, any
The p* would be extended to 2 more of those 7. Wouldn't make sense for
prange, pall or pany. So, just psum and pprod. ^, %%, %/%, &, and | are
listed in section 2, Group "ops" and seem different to sum and prod in
that sense.
Is there a case to add psum? Or have I missed something.
If psum, then why not pdiff (-), pprod (*) and precip (/) ? And
similarly, what about equivalent functions for ^, %%, %/%, &, and | ?
Hadley
--
RStudio / Rice University
http://had.co.nz/