lottery crawler

library(pipeR)
library(httr)
library(xml2)
library(rvest)
library(stringi)
library(stringr)

# windows才需要先把locale改成C，並用stringi::stri_conv轉成big5
if (Sys.info()["sysname"] == "Windows") {
  backupLocale <- Sys.getlocale("LC_COLLATE")
  Sys.setlocale("LC_ALL", 'C')
}

url <- "http://www.lotto-8.com/listlto539.asp?indexpage=%i&orderby=new"
maxPage <- GET(sprintf(url, 1)) %>>% content %>>% xml_find_all("//td/a[@target][3]") %>>%
  xml_attr("href") %>>% str_extract("=\\d+") %>>% str_sub(2L) %>>% as.integer

out <- lapply(1:maxPage, function(i){
  tmp <- GET(sprintf(url, i)) %>>% content %>>%
    xml_find_first("//table[@class='auto-style4']") %>>%
    html_table(header = TRUE)
  if (Sys.info()["sysname"] == "Windows")
    names(tmp) <- stri_conv(names(tmp), "UTF-8", "big5")
  tmp[[2]] <- iconv(tmp[[2]], "latin1", "ASCII", sub="")
  return(tmp[1:2])
}) %>>% do.call(what = rbind)

if (Sys.info()["sysname"] == "Windows") {
  Sys.setlocale(locale = backupLocale)
}