A line by line stye of reading in the hurricane data

read_hurricanes.R

library(tidyverse)
library(xml2)

url <- "http://www.aoml.noaa.gov/hrd/hurdat/hurdat2-nepac.html"

# Import ------------------------------------------------------------------

hurricanes <- read_html(url) %>%
  xml_find_first(".//pre") %>% 
  xml_text() %>% 
  write_file("hurricanes.csv")

hurricanes_file <- file("hurricanes.csv", open = "r")
readLines(hurricanes_file, n = 1) # first line is empty

# Something to hold data, don't know before reading how 
# long this should but it's less than total number
# of lines
data_blocks <- vector("list", 2000)
block <- 1
line <- readLines(hurricanes_file, n = 1)

while(length(line) > 0){
  # parse header line
  header <- scan(text = line, 
    what = list(
      id = character(), 
      name = character(), 
      nrows = integer()),
    sep = ",")
  
  # get corresponding data lines
  data_lines <- readLines(hurricanes_file, 
    n = header$nrows)
  
  # parse data lines
  data <- read.csv(text = data_lines, header = FALSE,
    stringsAsFactors = FALSE)
  data$hurricane <- header$id
  data$name <- header$name
  
  data_blocks[[block]] <- data
  
  # increment 
  line <- readLines(hurricanes_file, n = 1)
  block <- block + 1
}

close(hurricanes_file)
hurricanes <- do.call(rbind, data_blocks)