how to stem both words of bigrams about quanteda HOT 5 CLOSED

The workaround using tm+quanteda:
vs <- VectorSource(texts)
corp <- Corpus(vs)
corp <- tm_map(corp, removeWords, stopwords("english"))
corp <- tm_map(corp, stemDocument)
corp2 <- corpus(corp)
dtm <- dfm(corp2, clean=TRUE, bigrams=TRUE, stem=FALSE)

from quanteda.

Fixed in 0.8.7-8.

> dfm("Banking industry", stem = TRUE, ngrams = 2, verbose = FALSE)
Document-feature matrix of: 1 document, 1 feature.
1 x 1 sparse Matrix of class "dfmSparse"
       features
docs    bank_industri
  text1             1
> wordstem(tokenize("Banking industry", n = 2))
tokenizedText object from 1 document.
Component 1 :
[1] "Bank_industri"

from quanteda.

How does the stem completion works in quanteda, sorry if the question looks too basic but I couldn't find the answer so I decided to use the last resort (Ask the developer).

from quanteda.

You can do this by stemming words before making ngrams:

toks <- tokens(data_corpus_inaugural)
toks_stem <- tokens_wordstem(toks)
toks_stem_bigram <- tokens_ngrams(toks_stem)

from quanteda.

@atheeralattar stem completion is not implemented in quanteda. What's the use case for this? I've always been curious as to why people need this.

from quanteda.