[R] Error in "plot(aov.object)" after upgradation to R-4.0.0 and R-4.0.1 for given R-Script-Example

[peter dalgaard]

This is fallout from the stringsAsFactors changes. You have 'Permanganate' as a character vector and it runs afoul of this code

    aterms <- attributes(terms(x))
    dcl <- aterms$dataClasses[-aterms$response]
    facvars <- names(dcl)[dcl %in% c("factor", "ordered")]

which does not include 'Permanganate', even though the model fit has de facto promoted it to a factor. In the end, you try to multiply a px2-matrix by a 3-vector and things go poof.

This probably counts a bug in R, but I see that your code actually tries to pre-convert the variables to factors. However, you misspelled "Permangate"....

-pd

> On 19 Jun 2020, at 17:13 , Akhilesh Singh <akhileshsingh.igkv using gmail.com> wrote:
> 
> Dear learned experts of R,
> 
> I was writing a book through RStudio-Rmarkdown and had finally compiled it
> successfully based on R package R-3.6.2. Afterwards, I updated my R-3.6.2
> to R-4.0.0 and even later to R-4.0.1.
> 
> Then, the publishers demanded to recompile the book with font embedding, so
> I tried to recompile the book, when I found the following error:
> 
> "Error in (dm - 1) %*% ff : non-conformable arguments"
> 
> For convenience and reproducibility of the error, I am giving below the
> same code chunk as an R-Script-Example, wherein the error is occurring in
> the plot() function with the input of an aov() object.
> 
> R-Script-Example producing error:
> ==========================
> 
> setwd("E:/AKS-DATA-New/Software/R and allied
> packages/R-Markdown/knitr/MyBooks/STAT-512_STAT-564")
> getwd()
> 
> #After Upgrading to R-4.0.0 and even in R-4.0.1 following error in "plot()"
> function occurs:
> 
> Block=c(1,1,1,1,2,2,2,2,3,3,3,3)
> Permanganate=c("without","without","with","with","without","without","with","with","without","without","with","with")
> Sample.Size=c(0.25,1,0.25,1,0.25,1,0.25,1,0.25,1,0.25,1)
> Riboflavin=c(39.5,38.6,27.2,24.6,43.1,39.5,23.2,24.2,45.2,33,24.8,22.2)
> 
> #Creating data frame
> sned.2x2.woint=data.frame(Block, Sample.Size, Permanganate, Riboflavin)
> 
> #Declaring Block, Sample.Size, Permanganate as factors
> sned.2x2.woint$Block = factor(sned.2x2.woint$Block)
> sned.2x2.woint$Sample.Size = factor(sned.2x2.woint$Sample.Size)
> sned.2x2.woint$Permangate=factor(sned.2x2.woint$Permanganate)
> 
> #ANOVA of RBD when Block, Sample.Size, Permanganate are a fixed effects
> sned.2x2.woint.aov1=aov(Riboflavin ~ Block + Sample.Size + Permanganate +
> Sample.Size:Permanganate,data=sned.2x2.woint)
> 
> cat("ANOVA of RBD when Block, Sample.Size and Permanganate are fixed
> effects:\n")
> summary(sned.2x2.woint.aov1)
> 
> #ANOVA of RBD when Block, Sample.Size, Permanganate are fixed effects
> sned.2x2.woint.aov2=aov(Riboflavin ~ Block + Sample.Size + Permanganate,
> data=sned.2x2.woint)
> 
> cat("ANOVA of RBD when Block, Sample.Size and Permanganate are fixed
> effects:\n")
> 
> summary(sned.2x2.woint.aov2)
> 
> plot(sned.2x2.woint.aov2, which=1) #OK
> plot(sned.2x2.woint.aov2, which=2) #Ok
> plot(sned.2x2.woint.aov2, which=3) #OK
> plot(sned.2x2.woint.aov2, which=4) #OK
> 
> plot(sned.2x2.woint.aov2, which=5) #Error in (dm - 1) %*% ff :
> non-conformable arguments
> 
> plot(sned.2x2.woint.aov2, which=6) #OK
> =================================
> 
> I request the esteemed and learned experts of R to kindly me out to
> overcome this error.
> 
> With regards
> 
> 
> 
> -- 
> Dr. A.K. Singh
> Professor and Head (Agricultural Statistics)
> Department of Agricultural Statistics and Social Science (L)
> Indira Gandhi Krishi Vishwavidyalaya, Raipur-492 012,
> Chhattisgarh, India
> Mobile: +918770625795
> Email: akhileshsingh.igkv using gmail.com
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help using r-project.org mailing list -- To UNSUBSCRIBE and more, see
> 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.

-- 
Peter Dalgaard, Professor,
Center for Statistics, Copenhagen Business School
Solbjerg Plads 3, 2000 Frederiksberg, Denmark
Phone: (+45)38153501
Office: A 4.23
Email: pd.mes using cbs.dk  Priv: PDalgd using gmail.com