CODEWITHSUNDEEP
rurlencodingrcurlhttr
KERO
KERO

Reputation: 705

Avoid url encoding in R

I'm trying to get the lat/lon from the Google Geocoding API, but the request fails when the danish local characters are in the address. I suspect it's because the httr::GET function encodes the url, but I'm not really sure if I'm right.

If you copy/paste this link directly into your browser you get a valid result: http://maps.googleapis.com/maps/api/geocode/json?address=Søholmen+9,+4500+Denmark

But the code below is invalid even though the url is the same before it's parsed into the GET function. It works if I use an address without my local characters.

library(httr)
library(jsonlite)
library(stringr)

address <- "Søholmen 9, 4500 Denmark"
# address <- "Kronprinsesse Sofies Vej 6, 2000 Denmark"

base_url <- "http://maps.googleapis.com/maps/api/geocode/json?"

# An address OR components
geo_url <- paste0(base_url, "address=", str_replace_all(address, pattern = " ", replacement = "+"))

# Get the result
# get the content
# Parse the JSON
temp_geo_results <- httr::GET(url = URLencode(URL = geo_url), verbose())
temp_geo_results <- httr::content(temp_geo_results, as = "text")
temp_geo_results <- jsonlite::fromJSON(temp_geo_results)

Here's my sessionInfo()

R version 3.1.2 (2014-10-31)
Platform: x86_64-w64-mingw32/x64 (64-bit)

locale:
[1] LC_COLLATE=Danish_Denmark.1252  LC_CTYPE=Danish_Denmark.1252        LC_MONETARY=Danish_Denmark.1252
[4] LC_NUMERIC=C                    LC_TIME=Danish_Denmark.1252    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] stringr_0.6.2   jsonlite_0.9.10 httr_0.5       

loaded via a namespace (and not attached):
[1] RCurl_1.95-4.3 tools_3.1.2

EDIT: I removed a line of code not necessary for the question and added my sessionInfo.

Upvotes: 3

Views: 1323

Answers (4)

Andr&#233; Jakob
Andr&#233; Jakob

Reputation: 1

How I solved a similar problem: Setting Encoding between rawToChar and fromJSON like below (non executable).

library(httr)
library(jsonlite)

call_api <- GET("YOUR_URL",
                  add_headers(.headers=c(`Authorization` = "YOUR_KEY")))

strange_characters <- rawToChar(call_api$content) #wherever the raw_data is

# if you pass Encoding(strange_characters) you will get "unknown". So run the line below.
Encoding(strange_characters) <- "UTF-8" 
  
right_characters <- fromJSON(strange_characters)

Upvotes: 0

dimitris_ps
dimitris_ps

Reputation: 5951

You can use the rvest packages

library(rvest); library(jsonlite)
address <- "Søholmen 9, 4500 Denmark"
# address <- "Kronprinsesse Sofies Vej 6, 2000 Denmark"
base_url <- "http://maps.googleapis.com/maps/api/geocode/json?"

# An address OR components
geo_url <- paste0(base_url, "address=", str_replace_all(address, pattern = " ", replacement = "+"))
geo_url <- iconv(geo_url, to="UTF-8")

temp_geo_results <- html_text(html_nodes(html(geo_url) , "p"))
temp_geo_results <- fromJSON(temp_geo_results)

Upvotes: 0

jbaums
jbaums

Reputation: 27388

This seems to be an encoding issue.

The following works fine for me:

address <- "Søholmen 9, 4500 Denmark"
u <- sprintf("http://maps.googleapis.com/maps/api/geocode/json?address=%s", 
             gsub('\\s+', '+', enc2utf8(address)))

fromJSON(content(GET(u), as='text'))

Upvotes: 4

statespace
statespace

Reputation: 1664

I can share the crude way how I solved the very same problem in my language:

deencode <- function(text){
  output <- NULL
  for(i in 1:length(text)){
    temp <- text[i]
    temp <- gsub("ā", "a", temp)  
    temp <- gsub("Ā", "A", temp)
    temp <- gsub("č", "c", temp)
    temp <- gsub("Č", "C", temp)
    temp <- gsub("ē", "e", temp)
    temp <- gsub("Ē", "E", temp)
    temp <- gsub("ģ", "g", temp)
    temp <- gsub("Ģ", "G", temp)
    temp <- gsub("ī", "i", temp)
    temp <- gsub("Ī", "I", temp)
    temp <- gsub("ķ", "k", temp)
    temp <- gsub("Ķ", "K", temp)
    temp <- gsub("ļ", "l", temp)
    temp <- gsub("Ļ", "L", temp)
    temp <- gsub("ņ", "n", temp)
    temp <- gsub("Ņ", "N", temp)
    temp <- gsub("š", "s", temp)
    temp <- gsub("Š", "S", temp)
    temp <- gsub("ū", "u", temp)
    temp <- gsub("Ū", "u", temp)
    temp <- gsub("ž", "z", temp)
    temp <- gsub("Ž", "Z", temp)
    output <- c(output, temp)
  }
  return(output)
}

After this simplistic substitution it all worked, at least in Google geocode API.

Upvotes: -1

Related Questions