[R] odd behavior within R2HTML

[Milan Bouchet-Valat]

Le vendredi 15 février 2013 à 20:52 -0600, Erin Hodgess a écrit :
> Dear R People:
> 
> I'm using R2HTML but having a strange result.
> 
> Here is the original data:
> 
> resp trt block
> 90.3 A I
> 89.2 A II
> 98.2 A III
> 93.9 A IV
> 87.4 A V
> 97.9 A VI
> 92.5 B I
> 89.5 B II
> 90.6 B III
> 94.7 B IV
> 87.0 B V
> 95.8 B VI
> 85.5 C I
> 90.8 C II
> 89.6 C III
> 86.2 C IV
> 88.0 C V
> 93.4 C VI
> 82.5 D I
> 89.5 D II
> 85.6 D III
> 87.4 D IV
> 78.9 D V
> 90.7 D VI
> 
> And here are the commands:
> > resin1.df <- read.table("resin1.txt",header=TRUE)
> > #set up R-B anova
> > resin1.aov <- aov(resp ~ trt + block, data=resin1.df)
> > library(R2HTML
> + )
> > HTMLStart(outdir=getwd(),file="resin2",echo=TRUE)
> 
>  *** Output redirected to directory:  c:/R64/R-2.15.2/bin/x64
>  *** Use HTMLStop() to end redirection.[1] TRUE
> HTML> library(effects)
> Loading required package: lattice
> Loading required package: grid
> Loading required package: MASS
> Loading required package: nnet
> Loading required package: colorspace
> 
> Attaching package: ‘effects’
> 
> The following object(s) are masked from ‘package:datasets’:
> 
>     Titanic
> 
> HTML> summary(allEffects(resin1.aov)
> + )
>  model: resp ~ trt + block
> 
>  trt effect
> trt
>        A        B        C        D
> 92.81667 91.68333 88.91667 85.76667
> 
>  Lower 95 Percent Confidence Limits
> trt
>        A        B        C        D
> 90.46148 89.32815 86.56148 83.41148
> 
>  Upper 95 Percent Confidence Limits
> trt
>        A        B        C        D
> 95.17185 94.03852 91.27185 88.12185
> 
>  block effect
> block
>      I     II    III     IV      V     VI
> 87.700 89.750 91.000 90.550 85.325 94.450
> 
>  Lower 95 Percent Confidence Limits
> block
>       I      II     III      IV       V      VI
> 84.8155 86.8655 88.1155 87.6655 82.4405 91.5655
> 
>  Upper 95 Percent Confidence Limits
> block
>       I      II     III      IV       V      VI
> 90.5845 92.6345 93.8845 93.4345 88.2095 97.3345
> HTML> HTMLStop()
> [1] "c:/R64/R-2.15.2/bin/x64/resin2_main.html"
> > file.show("resin2_main.html")
> >
> 
> When I look in the resin2_main.html file, the results from the summary
> are not there.
> 
> Has anyone run into this before, please?
Yes, this is fully reproducible with examples from ?allEffects. The
problem is that summary.efflist() does not return an object that R2HTML
could print(). Instead, it calls print() directly on all of the separate
effects in the list, and returns invisible(NULL).

You can fix this by making summary.efflist() return an object containing
the needed data, and creating a function to print this object as below.
Maybe John Fox will want to apply this kind of change to the package.


Regards

summary.efflist <- function (object, ...) 
{
    form <- object[[1]]$formula
    attributes(form) <- NULL
    ret <- list(formula=form,
                eff=lapply(object, summary, ...))

    class(ret) <- "summary.efflist"
    ret
}

print.summary.efflist <- function(x, ...)
{
    cat(" model: ")
    print(x$formula)

    for (effect in names(x$eff)) {
        print(x$eff[[effect]], ...)
    }

    invisible(x)
}

> Thanks for any help!
> 
> Sincerely,
> Erin
> R-2.15.2, Windows, 64-bit
> 
>