Dictionary() is not supported anymore in tm package. How to emend code?

Question

I just noticed that after updating to tm v. 0.5-10 the function Dictionary() is not supported anymore. Is this an mistake? Or was it deprecated? Am

Answer 1

A more complete extract from the NEWS file:

\subsection{DEPRECATED & DEFUNCT}{
    \itemize{
      \item Following functions have been removed:
      \itemize{
        \item \code{Dictionary()} (use a character vector instead; use
          \code{Terms()} to extract terms from a document-term or term-document 
          matrix),

So. yes, it is deprecated and removed. And as Ben suggested, the authors intend you to use Terms(). Why you get the error is a matter only for idle speculation since you have not provided a data object and code the throws the error. One guess is that the object you are providing is not a TDM or DTM.

Answer 2

If you are using Dictionary as @Ben suggested, I think you could create a dummy function called Dictionary which just returned the character vector you passed to it.

Dictionary <- function(x) {
    if( is.character(x) ) {
        return (x)
    }
    stop('x is not a character vector')
}

However, longer term, it's probably better to roll up your sleeves and refactor the code.

Answer 3

As IShouldBuyABoat says, you haven't given us any clue about how you're using Dictionary so we can't really give you any specific answers (do update your question with more details).

In any case, the answer your question of 'how to update my code' is probably 'just delete Dictionary and it should be fine', as you can see here:

library(tm)
data(crude)

Find out what Dictionary did in earlier versions of the tm package:

methods(Dictionary)
getAnywhere(Dictionary.DocumentTermMatrix)
# function(x) structure(Terms(x), class = c("Dictionary", "character"))
getAnywhere(Dictionary.character)
# function (x)  structure(x, class = c("Dictionary", "character"))

Kind of a pointless function anyway, seems quite sensible to remove it. But how to update your code that depended on it?

You may have used Dictionary like this:

myDictionary <- Dictionary(c("some", "tokens", "that", "I", "am", "interested", "in"))
inspect(DocumentTermMatrix(crude, list(dictionary = myDictionary)))

Now that this function is not longer available, you'd do this instead, using a character vector: '

myTerms <- c("some", "tokens", "that", "I", "am", "interested", "in")
inspect(DocumentTermMatrix(crude, list(dictionary = myTerms)))

The output for these two examples is identical, first one was using tm version 0.5-9 and the second with version 0.5-10

The instruction in the NEWS to use Terms is if you want to get all the words in a document term matrix, like so

Terms(DocumentTermMatrix(crude))

If none of that helps you then you'll need to supply more detail about what you're trying to do.