R help - Jul 2010 - model.tables call fails with "Error in inherits(object, "formula")"

Hello,

I noticed that model.tables fails when applied to an aov() fit if called 
inside a function. The problem seems to occur when as.formula is used 
inside a function on a string containing
"<formula> + Error( x / y )"

The reason I tried to use as.formula is to generate dynamic calls to aov().
Here is a minimal example illustrating the problem:

## Example

test <- function(dat) {
     frmlb <- "yield ~ block";                    # WORKS
     aovfit <- aov(as.formula(frmlb), dat);
     print(model.tables(aovfit, "means"));

     frmlc <- "yield ~ block + Error(N/block)";   # DOES NOT WORK
     aovfit <- aov(as.formula(frmlc), dat);
     print(model.tables(aovfit, "means"));
}

utils::data(npk, package="MASS");
frmla <- "yield ~ block + Error(N/block)";       # WORKS
aovfit <- aov(as.formula(frmla), npk);
print(model.tables(aovfit, "means"));
test(npk);

## End of example


Output of sessionInfo():
R version 2.11.1 (2010-05-31)
i386-pc-mingw32

locale:
[1] LC_COLLATE=Finnish_Finland.1252  LC_CTYPE=Finnish_Finland.1252
[3] LC_MONETARY=Finnish_Finland.1252 LC_NUMERIC=C
[5] LC_TIME=Finnish_Finland.1252

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

loaded via a namespace (and not attached):
[1] tools_2.11.1

br Hannes G.

The error message actually is

Error in inherits(object, "formula") : object 'frmlc' not
found

and it is the omitted-by-you part after the colon that matters: the 
body of model.tables does not have 'frmlc' in its scope.

Here is one approach to using dynamic formulae that works in a lot of 
similar situations:

test <- function(dat) {
     frmlc <- as.formula("yield ~ block + Error(N/block)")
     aovfit <- eval(substitute(aov(form, dat),list(form=frmlc)))
     model.tables(aovfit, "means")
}

written without the pointless empty statements after the semicolons.


On Fri, 23 Jul 2010, Hannes wrote:
> Hello,
>
> I noticed that model.tables fails when applied to an aov() fit if called 
> inside a function. The problem seems to occur when as.formula is used
inside
> a function on a string containing
> "<formula> + Error( x / y )"
>
> The reason I tried to use as.formula is to generate dynamic calls to aov().
> Here is a minimal example illustrating the problem:
> ## Example
>
> test <- function(dat) {
>    frmlb <- "yield ~ block";                    # WORKS
>    aovfit <- aov(as.formula(frmlb), dat);
>    print(model.tables(aovfit, "means"));
>
>    frmlc <- "yield ~ block + Error(N/block)";   # DOES NOT
WORK
>    aovfit <- aov(as.formula(frmlc), dat);
>    print(model.tables(aovfit, "means"));
> }
>
> utils::data(npk, package="MASS");
> frmla <- "yield ~ block + Error(N/block)";       # WORKS
> aovfit <- aov(as.formula(frmla), npk);
> print(model.tables(aovfit, "means"));
> test(npk);
>
> ## End of example
>
>
> Output of sessionInfo():
> R version 2.11.1 (2010-05-31)
> i386-pc-mingw32
>
> locale:
> [1] LC_COLLATE=Finnish_Finland.1252  LC_CTYPE=Finnish_Finland.1252
> [3] LC_MONETARY=Finnish_Finland.1252 LC_NUMERIC=C
> [5] LC_TIME=Finnish_Finland.1252
>
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
>
> loaded via a namespace (and not attached):
> [1] tools_2.11.1
>
> br Hannes G.
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
-- 
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