For the record, many functions in numpy allow you to specify the axis over which reductions are to be performed--precisely what you're looking for. This sounds nice until you realize that the concept doesn't scale up to higher dimensions or multiple input piddles very nicely. This is why PDL introduces the axis manipulation methods and always operates on the left-most axes. It is far more general.
David On Tue, Jun 11, 2013 at 12:42 PM, Chris Marshall <[email protected]>wrote: > Hi Ingo- > > <replying to perldl list as this is not pdl-porters only> > > That is the current way to do so. While is is a lot > to write, it is very general. However, here are some > thoughts: > > - you could implement a new rotate() method that > does what you need > > - you could implement a do() method that evals and > generates the needed method > > - work for PDL3 is going to enable more flexible dimension > assignments and permutation for threading and argument > passing > > And yes, I find it tedious as well :-) > > --Chris > > On Tue, Jun 11, 2013 at 11:49 AM, Ingo Schmid <[email protected]> wrote: > > > > Hi, > > > > is there a simple way to specify a dimension for a given operation? What > I > > mean is for example rotate along the n-th dimension. I knwo that I can do > > > > $x->mv($n,0)->rotate($r)->mv(0,$n) > > > > but this is quite tedious, imo. > > > > I know there is PDL::Reduce for aggregates. > > > > What I mean is a syntax like > > > > $x->do('rotate',$n) > > > > or something similar. > > Does it exist? > > > > Ingo > > _______________________________________________ > Perldl mailing list > [email protected] > http://mailman.jach.hawaii.edu/mailman/listinfo/perldl > -- "Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it." -- Brian Kernighan
_______________________________________________ Perldl mailing list [email protected] http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
