Umlaut ¨ with package tm (text mining in R) - r

Question

I am trying to read some PDF docs using the package tm for text mining in R. However, my PDF are in german and I dont know how to deal with those Special characters.
library(tm)
pathname <- "J:/branchwarren/docs/tm/"
raw_corpus <- VCorpus(DirSource(directory=path,encoding="UTF-8"), readerControl=list(reader=readPDF,language="de"))
tdm <- TermDocumentMatrix(raw_corpus)
tdm_mat <- as.data.frame(tdm)
The Output tdm_mat for example is (where the columns are the frequencies in each PDF)
1 geschã¤ftsverlauf 9 9 1 3 0 0
2 gesellschaft 1 3 1 1 1 1
3 gesellschaft. 0 0 1 1 1 0
4 gesellschaftskapital 1 1 1 1 1 1
5 gestaltung 1 1 1 1 1 1
6 gesteigert 0 0 2 0 2 6
7 gesunden 0 1 0 1 1 1
8 gewinnreserve 1 1 1 1 1 1
9 gewinnverwendung) 1 1
As you notice, the character in the first row is not displayed correctly. It should be geschäftsverlauf.
Any help or suggestions? thanks in advance

Answer 1

Too long for a comment, but e.g. this works for me as expected:
library(tm)
dir.create(pathname <- tempfile())
writeLines("Der Geschäftsbericht war gut. Die Maßnahmen griffen.", tf <- tempfile(fileext = ".md"))
rmarkdown::render(input=tf, output_format="pdf_document", output_file="1.pdf", output_dir=pathname)
if(all(file.exists(Sys.which(c("pdfinfo", "pdftotext"))))) { # see ?readPDF
raw_corpus <- VCorpus(DirSource(directory=pathname, encoding="UTF-8"), readerControl=list(reader=readPDF,language="de"))
tdm <- TermDocumentMatrix(raw_corpus)
tdm_mat <- as.data.frame(as.matrix(tdm))
tdm_mat
}
# 1.pdf
# der 1
# die 1
# geschäftsbericht 1
# griffen. 1
# gut. 1
# maßnahmen 1
# war 1
My sessionInfo():
R version 3.3.0 (2016-05-03)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 7 x64 (build 7601) Service Pack 1
locale:
[1] LC_COLLATE=German_Germany.1252 LC_CTYPE=German_Germany.1252 LC_MONETARY=German_Germany.1252 LC_NUMERIC=C
[5] LC_TIME=German_Germany.1252
...
tm_0.6-2 NLP_0.1-8
...
Maybe an encoding mismatch? Try providing input data + your sessionInfo to debug & reproduce the error.