R help - Jan 2005 - integrate() and complex values

[apologies for possible multiple post]


Hi

The manpage for integrate() does not mention imaginary numbers.
I found the following unexpected:

 > integrate(function(x){0*x+1+1i},0,1)
1 with absolute error < 1.1e-14
 >

(I would expect 1+1i here)

One can write a little wrapper, but it's not straightforward.
Would it be hard to accommodate such functions in integrate()?

Some other functions returning complex values cause integrate() to trip 
up:

 > integrate(function(x){exp(1i*x)},0,1)
Error in integrate(function(x) { : maximum number of subdivisions 
reached


--
Robin Hankin
Uncertainty Analyst
Southampton Oceanography Centre
European Way, Southampton SO14 3ZH, UK
  tel  023-8059-7743

On Tue, 11 Jan 2005, Robin Hankin wrote:
> The manpage for integrate() does not mention imaginary numbers.
The help page does say

       f: an R function taking a numeric first argument and returning a
          numeric vector of the same length.

and in R complex numbers are not numeric (see ?numeric).  So you are
misusing it,
> I found the following unexpected:
>
>  > integrate(function(x){0*x+1+1i},0,1)
> 1 with absolute error < 1.1e-14
>  >
>
> (I would expect 1+1i here)

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595