[prev in list] [next in list] [prev in thread] [next in thread] 

List:       r-devel
Subject:    Re: [Rd] cumsum method in Math group
From:       Michael Lawrence <lawrence.michael () gene ! com>
Date:       2016-06-09 12:34:36
Message-ID: CAOQ5Nyf8iNTbd+aUrKnfF=ByzJWbcLJ3XoMa1grJyQ7qOmFQxg () mail ! gmail ! com
[Download RAW message or body]

On Thu, Jun 9, 2016 at 1:13 AM, Edzer Pebesma
<edzer.pebesma@uni-muenster.de> wrote:
> When running
>
> a = runif(10)
> class(a) = "foo"
> Math.foo = function(x, ...) {
>     NextMethod(.Generic)
> }
> signif(a, 3)
> cumsum(a)
>
>
> I don't understand why cumsum strips the class, but signif does not.
> Both claim in the documentation that "These are generic functions:
> methods can be defined for them individually or via the ‘Math' group
> generic."
>

There's no contract that the class be preserved. Only that dispatch
will happen. It's fairly unpredictable when R will preserve
attributes, like class. It's safest to cast the result in your Math
method. In general, who is to say whether the cumsum() of an object is
a valid instance of the same class? This also implies that, while
initially tempting, extending base types, in particular vector types,
will ultimately lead to unpredictable behavior and frustration.
Consider composition instead.

> --
> Edzer Pebesma
> Institute for Geoinformatics  (ifgi),  University of Münster
> Heisenbergstraße 2, 48149 Münster, Germany; +49 251 83 33081
> Journal of Statistical Software:   http://www.jstatsoft.org/
> Computers & Geosciences:   http://elsevier.com/locate/cageo/
> Spatial Statistics Society http://www.spatialstatistics.info
>
>
> ______________________________________________
> R-devel@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel

______________________________________________
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
[prev in list] [next in list] [prev in thread] [next in thread]