r - 如何在 R 中从 XML 中取消嵌套字典？

Question

我正在尝试将此 xml 转换为 R 中的数据框:

https://d396qusza40orc.cloudfront.net/getdata%2Fdata%2Frestaurants.xml

library(xml2)
library(tidyverse)

fileurl <- "https://d396qusza40orc.cloudfront.net/getdata%2Fdata%2Frestaurants.xml"
xmllist <- as_list(read_xml(fileurl)) 
xml_df = tibble::as_tibble(xmllist) %>%
        unnest_longer(response)

row_wider = xml_df %>%
        unnest_wider(response)

row_df = row_wider %>%
        unnest(cols = names(.)) %>%
        unnest(cols = names(.)) %>%
        readr::type_convert()

问题是“location_1”列是一个字典，当我取消嵌套时显示为 NA。如何将此字典的每个值放入此列中？非常感谢任何帮助，谢谢。

Answer 1

请求的地址数据以 JSON 形式存储在 XML 节点的属性中。
下面我提取属性，转换 JSON，然后合并。然后可以将生成的数据帧绑定(bind)到上面执行的工作。

详情请参阅评论。

library(xml2)
library(jsonlite)
library(tidyverse)

#read file as xml
page <- read_xml(fileurl)

#extract out the restaurant nodes into a vector
restaurants <- page %>% xml_find_all(".//row/row")
#get the address data which is stored as attribute data
addresses <- restaurants %>% xml_find_first(".//location_1") %>% xml_attr("human_address") 

#this is a vector of JSON data structures
#convert the JSON to a data frame

dfs <- lapply(addresses, function(address){
   address %>% fromJSON() %>% as.data.frame()
})
#combine all of the data frames
answer<- bind_rows(dfs)

answer

                     address      city state zip
1           4509 BELAIR ROAD Baltimore    MD    
2              1919 FLEET ST Baltimore    MD    
3             2844 HUDSON ST Baltimore    MD    
4            3998 ROLAND AVE Baltimore    MD    
5         2481 frederick ave Baltimore    MD    
6            2722 HARFORD RD Baltimore    MD