knitr, lavaan, and subscript out of bounds

Question

I am getting an error, Error in lines[[i]] : subscript out of bounds when trying to knit a chunk that reads in a model from an external file and then fits the model in lavaan.

I create the model in a .R file:

model <- readLines(con = textConnection('
  depression =~ thoughts + pain + brain + use + suic + talk + sitalone +         
                headaches + app + heart + cheek + cry + sleep + disob + cold +
                liedown + worries + alone + annoyed + holdhead + drinkal +
                insult + greet + think + mutter + trust + donoth + sad + bad +
                weak + notalk + forget + crycont + livedie
  '))

cat(model, file = 'scripts/mod.lav.f1.0', sep = '\n')

When running the .R file, it puts the following in the scripts/mod.lav.f1.0 file:

depression =~ thoughts + pain + brain + use + suic + talk + sitalone +         
              headaches + app + heart + cheek + cry + sleep + disob + cold +
              liedown + worries + alone + annoyed + holdhead + drinkal +
              insult + greet + think + mutter + trust + donoth + sad + bad +
              weak + notalk + forget + crycont + livedie

My .rnw file then reads in the scripts/mod.lav.f1.0 that specifies my model.

\documentclass{article}

\begin{document}

<<cfa, include=FALSE, tidy=FALSE>>=
# read in model from file
 mod.1f.0 <- readLines("scripts/mod.lav.f1.0")
# fit the model
 fit.1f.0 <- cfa(mod.1f.0, data = mydata, ordered=items)
@

\end{document}

The problematic statement in the chunk seems to be fit.1f.0 <- cfa(mod.1f.0, data = mydata, ordered=items). When knitting the document, I get an error: Error in lines[[i]] : subscript out of bounds.

I am able to run the chunk in R without any problems. The model is stored in mod.1f.0 and the fit is stored in fit.1f.0.

Any thoughts about what is causing this error?

Answer 1

Someone sent me another example that I was able to run, so I decided to delete my cache and all of the hidden project files that RStudio creates. It worked. Prior to doing this I had only tried restarting RStudio.