# COUNT THE NUMBER OF TIMES WORDS OCCUR IN A FILE.
#
# R. M. Neal, 2004.


# COUNT WORDS IN A FILE.  Takes the name of a file as an argument, defaulting
# to "", which means entering text by typing to the R console.  Returns a
# data frame with variables "word" and "count", where "word" contains the
# distinct words found in the file, in alphabetical order, and "count" contains 
# the count of how many times that word occurred (which will be at least 1, 
# since only words that occur are entered into the data frame).  
#
# Words are defined as a sequence of non-whitespace character separated by
# whitespace characters (space, newline, and tab).  This unfortunately means 
# that punctuation marks are considered part of a word.  (Actually, the presence
# of quote marks modifies this - it really does whatever the R "scan" procedure
# does as far as separating items goes.)
#
# An error is reported if the file is empty (has no words).

word.count <- function(file="")
{
  # Read all the words and sort them alphabetically.

  words.read <- sort(scan(file,what=""))

  if (length(words.read)<1)
  { stop("No words in file")
  }

  # Initialize vectors "word" and "count" to have just the first word, with
  # count of one.  Set "n" to the number of words found so far.

  word <- c(words.read[1])
  count <- c(1)
  n <- 1

  # Go through the remaining words, incrementing counts when the word is the
  # same as the previous one, and creating new word/count pairs when a new
  # word is encountered.

  if (length(words.read)>1)
  {
    for (i in 2:length(words.read))
    { 
      if (words.read[i]==word[n])
      { count[n] <- count[n] + 1
      }
      else
      { word <- c (word, words.read[i])
        count <- c (count, 1)
        n <- n + 1
      }
    }
  }

  # Put the word and count vectors together in a data frame to return.

  data.frame(word,count)
}