define subset argument for function lm as variable?

On Wed, Aug 29, 2012 at 3:56 AM, Milan Bouchet-Valat <nalimilan at club.fr> wrote:

Le mardi 21 ao?t 2012 ? 07:51 -0700, Joshua Wiley a ?crit :

Hi Rainer,

You could try:

subs <- expression(dead==FALSE & recTreat==FALSE)

lme(formula, subset = eval(subs))

Not tested, but something along those lines should work.

Out of curiosity, why isn't "subset" (and "weights", which is very
similar in that regard) evaluated in the "data" environment, just like
the formula? Is this for historical reasons, or are there drawbacks to
such a feature?

I am not sure about weights offhand, but subset is evaluated in the
data environment----that is why that solution works.  The original
question was how to setup the expression as an object that was passed
to subset.  The trick is to avoid having the logical expression
evaluated when the object is created, which I avoided by using
expression, and then in lme() forcing the evaluation of the object.

It seems very common to pass a data frame via the "data" argument, and
use variables from it for subsetting and/or weighting.


Regards

Josh

On Tue, Aug 21, 2012 at 7:44 AM, Rainer M Krug <r.m.krug at gmail.com> wrote:

Hi

I want to do a series of linear models, and would like to define the input
arguments for lm() as variables. I managed easily to define the formula
arguments in a variable, but I also would like to have the "subset" in a
variable. My reasoning is, that I have the subset in the results object.

So I wiould like to add a line like:

subs <- dead==FALSE & recTreat==FALSE

which obviously does not work as the expression is evaluated immediately. Is
is it possible to do what I want to do here, or do I have to go back to use

dat <- subset(dat, dead==FALSE & recTreat==FALSE)

?



dat <- loadSPECIES(SPECIES)
feff <- height~pHarv*year               # fixed effect in the model
reff <- ~year|plant                     # random effect in the model, where
year is the
dat.lme <- lme(
             fixed = feff,                           # fixed effect in the
model
             data  = dat,
             random = reff,                          # random effect in the
model
             correlation = corAR1(form=~year|plant), #
             subset = dead==FALSE & recTreat==FALSE, #
             na.action = na.omit
             )
dat.lm <- lm(
            formula =  feff,              # fixed effect in the model
            data = dat,
            subset = dead==FALSE & recTreat==FALSE,
            na.action = na.omit
            )

Thanks,

Rainer

--
Rainer M. Krug, PhD (Conservation Ecology, SUN), MSc (Conservation Biology,
UCT), Dipl. Phys. (Germany)

Centre of Excellence for Invasion Biology
Stellenbosch University
South Africa

Tel :       +33 - (0)9 53 10 27 44
Cell:       +33 - (0)6 85 62 59 98
Fax :       +33 - (0)9 58 10 27 44

Fax (D):    +49 - (0)3 21 21 25 22 44

email:      Rainer at krugs.de

Skype:      RMkrug