[R] Wrong environment when evaluating and expression?

[Joshua Wiley]

Thanks Gabor, that makes sense now.

In case anyone else runs into something similar, I ended up just
passing a character string of the formula so it could be coerced to a
formula in the correct environment.

Thanks again,

Josh

On Mon, Jul 4, 2011 at 4:26 AM, Gabor Grothendieck
<ggrothendieck at gmail.com> wrote:
> On Mon, Jul 4, 2011 at 4:11 AM, Joshua Wiley <jwiley.psych at gmail.com> wrote:
>> Hi All,
>>
>> I have constructed two expressions (e1 & e2).  I can see that they are
>> not identical, but I cannot figure out how they differ.
>>
>> ###############
>> dat <- mtcars
>> e1 <- expression(with(data = dat, lm(mpg ~ hp)))
>> e2 <- as.expression(substitute(with(data = dat, lm(f)), list(f = mpg ~ hp)))
>>
>> str(e1)
>> str(e2)
>> all.equal(e1, e2)
>> identical(e1, e2) # false
>>
>> eval(e1)
>> eval(e2)
>> ################
>>
>> The context is trying to use a list of formulae to generate several
>> models from a multiply imputed dataset.  The package I am using (mice)
>> has methods for with() and that is how I can (easily) get the pooled
>> results.  Passing the formula directly does not work, so I was trying
>> to generate the entire call and evaluate it as if I had typed it at
>> the console, but I am missing something (probably rather silly).
>>
>
> In e1, mpg ~ hp is a call object but in e2 its a formula with an environment:
>
>> e1[[1]][[3]][[2]]
> mpg ~ hp
>> e2[[1]][[3]][[2]]
> mpg ~ hp
>>
>> class(e1[[1]][[3]][[2]])
> [1] "call"
>> class(e2[[1]][[3]][[2]])
> [1] "formula"
>>
>> environment(e2[[1]][[3]][[2]])
> <environment: R_GlobalEnv>
>
> --
> Statistics & Software Consulting
> GKX Group, GKX Associates Inc.
> tel: 1-877-GKX-GROUP
> email: ggrothendieck at gmail.com
>



-- 
Joshua Wiley
Ph.D. Student, Health Psychology
University of California, Los Angeles
http://www.joshuawiley.com/