termdocumentmatrix raises error in R

Question

I have the following code:

library(tm)
text<-readLines("anyText.txt")
corpus<-Corpus(VectorSource(text))
corpus<-tm_map(corpus,content_transformer(tolower))
inspect(corpus)
corpus<-tm_map(corpus,removePunctuation)
stopwords<-c(stopwords('english'),"available","via")
corpus<-tm_map(corpus,removeWords,stopwords)
tempCorpus<-corpus
inspect(tempCorpus)
library(ctv)
library(SnowballC)
corpus<-tm_map(corpus,stemDocument)
inspect(corpus)
corpusT<- tm_map(corpus, PlainTextDocument)
corpusT<-tm_map(corpusT,stemCompletion,dictionary=tempCorpus)
dtm<-TermDocumentMatrix(corpusT,control=list(minWordLength=1))

but I got the error:

   Error: inherits(doc, "TextDocument") is not TRUE

I found that when I comment the line:

corpusT<-tm_map(corpusT,stemCompletion,dictionary=tempCorpus)

the program works fine, but for what I know that last line is for updating the list of steem words with the ones I got in tempCorpus so I need it.

How can I correct that error?

Possible duplicate of [DocumentTermMatrix error on Corpus argument](http://stackoverflow.com/questions/24191728/documenttermmatrix-error-on-corpus-argument) — Hardik Gupta, Jan 12 '17 at 12:31

score 2 · Answer 1 · edited May 23 '17 at 12:24

2

The code corpusT<-tm_map(corpusT,stemCompletion,dictionary=tempCorpus) is not returning the TextDocuments so you are getting an error.

Adding the code corpusT<- tm_map(corpusT, PlainTextDocument) before creating dtm and after stemCompletion should fix the issue.

Your last part of the code should look as below:

inspect(corpus)
corpusT<-tm_map(corpus,stemCompletion,dictionary=tempCorpus)
corpusT<- tm_map(corpusT, PlainTextDocument)
dtm<-TermDocumentMatrix(corpusT,control=list(minWordLength=1))

For additional info please refer, https://stackoverflow.com/a/24206825/3858156

edited May 23 '17 at 12:24

Community

1
1

answered Mar 31 '15 at 03:24

Manohar Swamynathan

2,065
21
23

your solution is wrong and it has been posted before the issue, when one uses PlainTextDocument it erases the metadata – Little Mar 31 '15 at 03:26
Sorry, I have edited the answer. I meant `PlainTextDocument` should be run after `stemCompletion` and before `dtm`. Does that help? – Manohar Swamynathan Mar 31 '15 at 03:41
1

Thanks, but the error is still there, the metadata gets deleted – Little Mar 31 '15 at 03:54
Welcome, is it possible to give more information on meta data getting deleted and what exactly you are looking for in meta data? – Manohar Swamynathan Mar 31 '15 at 04:12

score 0 · Answer 2 · answered Dec 22 '15 at 11:23

0

I have found that I need to wrap all functions passed into tm_map in content_transformer, as you did with tolower. Then the error goes away.

answered Dec 22 '15 at 11:23

dnuttle

3,810
2
19
19

Not suggesting something like: `corpusT <- tm_map(corpusT, content_transformer(PlainTextDocument))`. That doesn't work. – CodeMonkey Jul 26 '16 at 16:42

termdocumentmatrix raises error in R

2 Answers2