draft/getSunnybank.R

In leoluyi/CRMaddress: Crawling address in Taiwan

getSunnybank <- function(){
  mylist <- list()
  for (i in 1:3)
  {
    
    url<- sprintf("https://www.sunnybank.com.tw/Branches.asp?page=%s&lmenuid=5&smenuid=99&areaid=1", i)
    res <- GET(url, encoding='big5', user_agent('Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/43.0.2357.130 Safari/537.36'),
               set_cookies('cookiesession1'='5628CEA7YN4YGJINJSOUNT2VFTEG1965'))
    res2 <- content(res,"text",encoding='big5')
    res2 <- toUTF8(res2)
    html <- htmlParse(res2, encoding = "utf8")
    
    tables <- readHTMLTable(html)
    length(tables)
    
    data <-tables[[1]]
    data1 <- data[,-6]
    data2 <- na.exclude(data1)
    data3 <- data2[-1,]
    data4 <- data3[-1,]
    data5 <- data4[,-1]
    
    addr  <- str_extract_all(unlist(data5[,2]), ".+(號|樓|F)")
    tel_1 <- unlist(data5[,3])
    fax_1 <- unlist(data5[,4])
    
    data05=data.frame('陽信銀行', data5[,1], unlist(tel_1), unlist(fax_1), unlist(addr), stringAsFactors=FALSE) 
    data06 <- data05[,-6]
    
    data_sum <- data06
    
    mylist=rbind(mylist,data_sum)
    
  }
  
  url_A <-c("https://www.sunnybank.com.tw/Branches.asp?lmenuid=5&smenuid=99&areaid=2",
            "https://www.sunnybank.com.tw/Branches.asp?lmenuid=5&smenuid=99&areaid=4",
            "https://www.sunnybank.com.tw/Branches.asp?lmenuid=5&smenuid=99&areaid=5")
  
  mylist_1 <- list()
  for (i in 1:3)
  {
    url <- url_A[i]
    res <- GET(url, encoding='big5', user_agent('Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/43.0.2357.130 Safari/537.36'),
               set_cookies('cookiesession1'='5628CEA7YN4YGJINJSOUNT2VFTEG1965'))
    res2 <- content(res,"text",encoding='big5')
    res2 <- toUTF8(res2)
    html <- htmlParse(res2, encoding = "utf8")
    
    tables <- readHTMLTable(html)
    length(tables)
    
    data <-tables[[1]]
    data1 <- data[,-6]
    data2 <- na.exclude(data1)
    data3 <- data2[-1,]
    data4 <- data3[-1,]
    data5 <- data4[,-1]
    
    addr  <- str_extract_all(unlist(data5[,2]), ".+(號|樓|F)")
    tel_1 <- unlist(data5[,3])
    fax_1 <- unlist(data5[,4])
    
    data05=data.frame('陽信銀行', data5[,1], unlist(tel_1), unlist(fax_1), unlist(addr), stringAsFactors=FALSE) 
    data06 <- data05[,-6]
    
    data_sum <- data06
    mylist_1=rbind(mylist_1,data_sum)
  }
  
  data_new = rbind(mylist, mylist_1)
  
  colnames(data_new) = c("brand_nm", "store_nm", "tel_no", "fax_no", "addr")
data_new
}