Error in R Markdown: Data must be a character vector

I am trying to run code in markdown in RStudio when knitting to html. The code runs without error as an R script. I am reading in files as a corpus using the tm text mining package. With each file (with each element in the corpus), I am processing it with code like the following, but I am getting an error apparently with this statement which uses a function from the syuzhet package


The error looks like this

Error in get_nrc_sentiment(AttdAI) : Data must be a character vector.
Calls: <Anonymous> ... withCallingHandlers -> withVisible -> eval -> eval -> get_nrc_senti

Here is the relevant code

for (cor in 1:length(my_corpus)){
  Essay_Corpus <- Corpus(VectorSource(my_corpus[cor]))

  # Convert the text to lower case
  Essay_Corpus <- tm_map(Essay_Corpus, content_transformer(tolower))
  # Remove numbers
  Essay_Corpus <- tm_map(Essay_Corpus, removeNumbers)
  # Remove english common stopwords
  Essay_Corpus <- tm_map(Essay_Corpus, removeWords, stopwords("en"))
  # Remove punctuations
  Essay_Corpus <- tm_map(Essay_Corpus, removePunctuation)
  # Eliminate extra white spaces
  Essay_Corpus <- tm_map(Essay_Corpus, stripWhitespace)
  # Text stemming
  Essay_Corpus <- tm_map(Essay_Corpus, stemDocument)
  # Remove additional stopwords
  Essay_Corpus <- tm_map(Essay_Corpus, removeWords, c("also", "can", "may", "even", "will", "however", "like", "many","retrieved","like" ,"name","data","ghotbi"))
  #Create term document matrix
  dtm_essay <- DocumentTermMatrix(Essay_Corpus)
  tdm_essay <- TermDocumentMatrix(Essay_Corpus)
  m_essay <- as.matrix(tdm_essay)
  v_essay <- sort(rowSums(m_essay),decreasing=TRUE)
  d_essay <- data.frame(word = names(v_essay),freq=v_essay)
  findAssocs(tdm_essay, c("vaccine", "time"), c(0))
  findAssocs(tdm_essay, c("job", "human"), c(0))
  #Draw word cloud

  Essay_wc[[cor]] <- ggplot(d_essay, aes(label = word, size = freq)) + geom_text_wordcloud()
  #Sentiment analysis
  AttdAI <-[cor])
  td_new <- data.frame(rowSums(td[1:2]))
  #Transformation and cleaning
  names(td_new)[1] <- "count"
  td_new <- cbind("sentiment"= rownames(td_new), td_new)
  allEssaysSentsAfter <- cbind(allEssaysSentsAfter,td_new[,2])
  rownames(td_new) <- NULL
  td_emo <-td_new[1:8,]
  #Calculate difference in Emotion
  afterSentiment <- cbind(afterSentiment,as.matrix(td_new[,2]))
  emo_plot[[cor]] <- qplot(sentiment, xlab="After Essay emotions", data=td_emo, weight=count, geom="bar",fill=sentiment)+ggtitle(pdfNames[cor])+theme(axis.title.x = element_text(size = 9, lineheight = .9,family = "Times", face = "bold.italic", colour = "red")) + theme(plot.title = element_text(size = 9, lineheight = .9,family = "Times", face = "bold.italic", colour = "red"))+ylim(0,22)
  sentiment_plot[[cor]] <- qplot(sentiment, xlab ="After Essay sentiments in binary terms", data=td_sentiment, weight=count, geom="bar",fill=sentiment)+ggtitle(pdfNames[cor])+theme(axis.title.x = element_text(size = 9, lineheight = .9,family = "Times", face = "bold.italic", colour = "red")) + theme(plot.title = element_text(size = 9, lineheight = .9,family = "Times", face = "bold.italic", colour = "red"))+ylim(0,55)

  coll2 <- textstat_collocations(AttdAI, size = 4:6)
  freqterms <- findFreqTerms(dtm_essay, lowfreq = 50)

Can anyone tell me how I can run this in markdown?

For reference, the libraries being used are

library(readtext)         # To read .txt files
library(stm)              # For structural topic models
library(stminsights)      # For visual exploration of STM
library(gsl)              # Required for the topicmodels package
library(topicmodels)      # For topicmodels
library(caret)            # For machine learning


is a syuzhet function with a signature

get_nrc_sentiment(char_v, cl = NULL, language = "english", lowercase = TRUE)

where char_v is a character vector, e.g.

c("alpha", "beta")

The function is being given as char_v AttdAI,which is the return value of from the textreg package. As noted in it's documentation

sometimes you need to convert your tm object to a string vector for various reasons, the main one being handing it to the C++ method. It is ugly, \dots
It is therefore a possibly better decision to pass a filename to a plain-text file to the textreg call to be loaded by C++ directly. See textreg

which indicates that get_nrc_sentiment may fail, there is an alternative and suggests that it is not worthwhile attempting to find why it might work in the console but not in the rendering.

Please see the FAQ on cross posting.

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.