R help - Jun 2005 - abbreviate

> p = data.frame(high=c(5,2), settle=c(3,4))
> p
  high settle
1    5      3
2    2      4

What is the most abbreviated way to apply:
if (p$high < p$settle) p$high = p$settle

I want to modify p to become:
> p
  high settle
1    5      3
2    4      4

On Wed, 15 Jun 2005, Omar Lakkis wrote:
>> p = data.frame(high=c(5,2), settle=c(3,4))
>> p
>  high settle
> 1    5      3
> 2    2      4
>
> What is the most abbreviated way to apply:
> if (p$high < p$settle) p$high = p$settle
>
> I want to modify p to become:
>> p
>  high settle
> 1    5      3
> 2    4      4
p[[1]] <- pmax(p[[1]], p[[2]])

seems to need a rather small number of keystrokes at the expense of 
readability (I would otherwise use p$high etc).

-- 
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

On Wed, 2005-06-15 at 13:59 -0400, Omar Lakkis wrote:
> > p = data.frame(high=c(5,2), settle=c(3,4))
> > p
>   high settle
> 1    5      3
> 2    2      4
> 
> What is the most abbreviated way to apply:
> if (p$high < p$settle) p$high = p$settle
> 
> I want to modify p to become:
> > p
>   high settle
> 1    5      3
> 2    4      4
Probably the easiest would be:
> p
  high settle
1    5      3
2    2      4
> p$high <- with(p, ifelse(high < settle, settle, high))
> p
  high settle
1    5      3
2    4      4


See ?ifelse and ?with

HTH,

Marc Schwartz

On Wed, 2005-06-15 at 19:33 +0100, Prof Brian Ripley
wrote:
> On Wed, 15 Jun 2005, Omar Lakkis wrote:
> 
> >> p = data.frame(high=c(5,2), settle=c(3,4))
> >> p
> >  high settle
> > 1    5      3
> > 2    2      4
> >
> > What is the most abbreviated way to apply:
> > if (p$high < p$settle) p$high = p$settle
> >
> > I want to modify p to become:
> >> p
> >  high settle
> > 1    5      3
> > 2    4      4
> 
> p[[1]] <- pmax(p[[1]], p[[2]])
> 
> seems to need a rather small number of keystrokes at the expense of 
> readability (I would otherwise use p$high etc).
I do like that approach. Definitely less keystrokes...

:-)

Marc