split-narrative-file.R

# source me and then use:
#
#   splitfile("workbook.xlsx", 1)
#
# to extract sheet 1 from workbook.xlsx

# set these if they vary
# subj_col determines the name of the output file
subj_col <- 1
# content_col determines the column with the contents of the file
content_col <- 2
# is there a header row?
has_header <- TRUE
# where to write them out?
output_directory <- "txt"

require('gdata')
splitfile <- function(filename, sheet = 1) {
  if (!file.exists(output_directory)) {
    dir.create(output_directory)
  }
  
  x <- read.xls(filename, sheet, header = has_header, quote = '', method='tab')
  
  if (subj_col > ncol(x) || content_col > ncol(x)) {
    error('Either subject col or content col are outside number of columns available')
  }
  
  n <- nrow(x)
  for (i in 1:n) {
    sn <- trim_quotes(x[i, subj_col])
    if (nchar(sn)>17) {
      # probably a bad conversion from numeric to string
      sn <- as.character(as.numeric(sn))
    }
    
    outfn <- sprintf('%s/%s.txt', output_directory, sn)
    print(outfn)
    fout <- file(outfn)
    writeLines(as.character(trim_quotes(x[i, content_col])), fout)
    close(fout)
  }
}

trim_quotes <- function(cell) {
  # remove the quotes from the first and last character
  x <- gsub('[(^")("$)]', '', cell)
  gsub('\\\\', '"', x)
}