R help - Jan 2005 - Name conflicts when passing arguments for one function to another

I am fairly new to R. I find it surprising that

f <- function(x,a) {x-a}
uniroot(f, c(0,1), a=.5)

works, but

integrate(f, 0, 1, a=.5)

gives an error: Error in integrate(f, 0, 1, a = 0.5) : argument 4
matches multiple formal arguments

What is the best way of avoiding such surprises? Is there a way of
telling integrate() that the 'a' argument is for f()?

If I wrote my own function along the lines of uniroot() or integrate()
is there a better way of passing on arguments?

-- 
Graham Jones, author of SharpEye Music Reader
http://www.visiv.co.uk
21e Balnakeil, Durness, Lairg, Sutherland, IV27 4PT, Scotland, UK

Graham Jones <maillists at visiv.co.uk> writes:
> I am fairly new to R. I find it surprising that
> 
> f <- function(x,a) {x-a}
> uniroot(f, c(0,1), a=.5)
> 
> works, but
> 
> integrate(f, 0, 1, a=.5)
> 
> gives an error: Error in integrate(f, 0, 1, a = 0.5) : argument 4
> matches multiple formal arguments
> 
> What is the best way of avoiding such surprises? Is there a way of
> telling integrate() that the 'a' argument is for f()?
> 
> If I wrote my own function along the lines of uniroot() or integrate()
> is there a better way of passing on arguments?
It's mainly a design issue. Had the "..." in the definition of
integrate been further to the left, then your problem wouldn't exist,
but you wouldn't be able to use abbreviations for abs.tol and friends. 

A simple (if tedious) workaround is

integrate(f, 0, 1, a=.5, abs.tol=.Machine$double.eps^0.25, aux=NULL)

a more thoroughly effective one could be

myintegrate <- function (f, lower, upper, ..., subdivisions = 100,
    rel.tol = .Machine$double.eps^0.25, abs.tol = rel.tol,
    stop.on.error = TRUE, keep.xy = FALSE, aux = NULL) 
 integrate(f, lower, upper, subdivisions, rel.tol, abs.tol, 
             stop.on.error, keep.xy, aux, ...)


-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)             FAX: (+45) 35327907

On Sat, 29 Jan 2005, Graham Jones wrote:
> I am fairly new to R. I find it surprising that
>
> f <- function(x,a) {x-a}
> uniroot(f, c(0,1), a=.5)
>
> works, but
>
> integrate(f, 0, 1, a=.5)
>
> gives an error: Error in integrate(f, 0, 1, a = 0.5) : argument 4
> matches multiple formal arguments
>
> What is the best way of avoiding such surprises? Is there a way of
> telling integrate() that the 'a' argument is for f()?
>
> If I wrote my own function along the lines of uniroot() or integrate()
> is there a better way of passing on arguments?
You are being caught by partial matching. Ideally integrate would have arg 
list

function (f, lower, upper, ..., subdivisions = 100,
     rel.tol = .Machine$double.eps^0.25,
     abs.tol = rel.tol, stop.on.error = TRUE, keep.xy = FALSE,
     aux = NULL)

(cf args(integrate)).  But the original author did not do it that way.
[The point is that args after ... have to be given in full.]

You can resolve this by explicitly giving abs.tol and aux:
>  integrate(f, 0, 1, abs.tol =1e-4, aux=NULL, a=.5)
1.758178e-18 with absolute error < 2.8e-15

or not using an argument name that has no meaning and is so subject to 
mis-matching.

-- 
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 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595