下载多个txt文件R

Question

I want to download a number of .txt-files. I have a data frame'"New_test in which the urls are under 'url' and the dest. names under 'code

"New_test.txt"

"url"   "code"
"1" "http://documents.worldbank.org/curated/en/704931468739539459/text/multi-page.txt" "704931468739539459.txt"
"2" "http://documents.worldbank.org/curated/en/239491468743788559/text/multi-page.txt"  "239491468743788559.txt"
"3" "http://documents.worldbank.org/curated/en/489381468771867920/text/multi-page.txt"  "489381468771867920.txt"
"4" "http://documents.worldbank.org/curated/en/663271468778456388/text/multi-page.txt"  "663271468778456388.txt"
"5" "http://documents.worldbank.org/curated/en/330661468742793711/text/multi-page.txt"  "330661468742793711.txt"
"6" "http://documents.worldbank.org/curated/en/120441468766519490/text/multi-page.txt"  "120441468766519490.txt"
"7" "http://documents.worldbank.org/curated/en/901481468770727038/text/multi-page.txt"  "901481468770727038.txt"
"8" "http://documents.worldbank.org/curated/en/172351468740162422/text/multi-page.txt"  "172351468740162422.txt"
"9" "http://documents.worldbank.org/curated/en/980401468740176249/text/multi-page.txt"  "980401468740176249.txt"
"10" "http://documents.worldbank.org/curated/en/166921468759906515/text/multi-page.txt" "166921468759906515.txt"
"11" "http://documents.worldbank.org/curated/en/681071468781809792/text/DRD169.txt" "681071468781809792.txt"
"12" "http://documents.worldbank.org/curated/en/358291468739333041/text/multi-page.txt" "358291468739333041.txt"
"13" "http://documents.worldbank.org/curated/en/716041468759870921/text/multi0page.txt" "716041468759870921.txt"
"14" "http://documents.worldbank.org/curated/en/961101468763752879/text/34896.txt"  "961101468763752879.txt"`

this is the script

rm(list=ls())

require(quanteda)
library(stringr)

workingdir <-setwd("~/Study/Master/Thesis/Mining/R/WorldBankDownl") 
test <- read.csv(paste0(workingdir,"/New_test.txt"), header = TRUE, 
stringsAsFactors = FALSE, sep="\t")

#Loop through every url in test_df and download in target directory with name = code
 for (url in test) {
 print(head(url))
 print(head(test$code))
 destfile <- paste0('~/Study/Master/Thesis/Mining/R/WorldBankDownl/Sources/', test$code)
 download.file(test$url, destfile,  method = "wget", quiet=TRUE)

And this is the error I get

Error in download.file(test$url, destfile, method = "wget", quiet = TRUE) : 
'url' must be a length-one character vector

Answer 1

Here's one way to do it that is a bit simpler. You will need to substitute your test$url for the txturls (both are character vectors with the URL of the text file).

txturls <- c("http://documents.worldbank.org/curated/en/704931468739539459/text/multi-page.txt", 
             "http://documents.worldbank.org/curated/en/239491468743788559/text/multi-page.txt",
             "http://documents.worldbank.org/curated/en/489381468771867920/text/multi-page.txt")

library("quanteda")

txt <- character()
for (i in txturls) {
    # read the file from the URL
    temp <- readLines(url("http://documents.worldbank.org/curated/en/704931468739539459/text/multi-page.txt"))
    # concatenate lines into one text
    temp <- texts(temp, groups = 1)
    # remove form feed character
    txt <- gsub("\\f", "", txt)
    # concatenate into the vector
    txt <- c(txt, temp)
}

# form the quanteda corpus
urlcorp <- corpus(txt, docvars = data.frame(source = txturls, stringsAsFactors = FALSE))
summary(urlcorp)
# Corpus consisting of 3 documents:
# 
#  Text Types Tokens Sentences                                                                           source
#     1  1343   5125       135 http://documents.worldbank.org/curated/en/704931468739539459/text/multi-page.txt
#   1.1  1343   5125       135 http://documents.worldbank.org/curated/en/239491468743788559/text/multi-page.txt
#   1.2  1343   5125       135 http://documents.worldbank.org/curated/en/489381468771867920/text/multi-page.txt
# 
# Source: /Users/kbenoit/Dropbox (Personal)/GitHub/dictionaries_paper/* on x86_64 by kbenoit
# Created: Tue Mar 20 15:51:05 2018
# Notes: 

Answer 2

Everyone, thank you for helping me. For me the solution was changing the method I used. 'wget' demanded 1 url in 'url' and the same for 'destfile'. 'A length-one character vector' Both 'url' and 'destfile' are length-fourteen character vectors. Now I use the method 'libcurl', which demands that the length of the character vectors is equal in 'url' and 'destfile'. If you use this method, make sure that quiet = TRUE.

Furthermore, it is possible that you have a working loop but that you get the error.

Error in download.file(test$url, destfile, method = "libcurl", quiet = TRUE) : 
cannot download any files
In addition: There were 50 or more warnings (use warnings() to see the first 50) 

This means that your source can't keep up with your calls, you're basically DDOS'ing, so the loop has to be slowed down.

rm(list=ls())

require(quanteda)
library(stringr)

workingdir <-setwd("~/Study/Master/Thesis/Mining/R/WorldBankDownl") 
test <- read.csv(paste0(workingdir,"/New_test.txt"), header = TRUE, 
stringsAsFactors = FALSE, sep="\t")

test <- data.frame(test)


#Loop through every url in comb_df and download in target directory with name = code, if you get an error and no files are downloaded try to slow down the loop.
for (url in test) {
 print(head(url))
 destfile <- paste0(workingdir, '/Sources/WB_', test$code)
 download.file(test$url, destfile, method = "libcurl", quiet = TRUE)
}