Here is the code. Meanwhile I've already detected the problem. Almost everytime I use LDA package with tm package. The latest version of tm package changed the corpus output. Previously meta data in corpus (tmp_map function) was present only at master level, now meta data is also present at a row level.
This is my solution doc =sapply(as.list(doc), function(x) (x[[1]])) before lexicalize(doc)
From Uros Godnov:
Here is the code. Meanwhile I've already detected the problem. Almost everytime I use LDA package with tm package. The latest version of tm package changed the corpus output. Previously meta data in corpus (tmp_map function) was present only at master level, now meta data is also present at a row level.
This is my solution doc =sapply(as.list(doc), function(x) (x[[1]])) before lexicalize(doc)