在R语言中通过httr包获取XML并转换为DataFrame的方法

Got it, let's break down how to take that XML response from your httr POST request and turn it into a clean DataFrame. I'll use xml2 for parsing XML (it's intuitive and plays well with tidyverse tools) and dplyr/purrr for wrangling the data into shape.

Step 1: Install & Load Required Packages #

First, make sure you have these packages installed (if not, run the install commands):

# Install packages if needed
install.packages(c("httr", "xml2", "dplyr", "purrr"))

# Load them
library(httr)
library(xml2)
library(dplyr)
library(purrr)

Step 2: Parse the XML Response #

Assuming you already sent your POST request and have the response stored in an object called resp, we'll parse it into an XML document:

# Send POST request (replace with your actual URL and body)
resp <- POST(
  url = "your-api-endpoint-url",
  body = list(your_parameters_here),
  encode = "form" # adjust encode type based on your API requirements
)

# Parse the response into an XML document
xml_doc <- content(resp, as = "parsed", type = "application/xml")

Step 3: Extract Terminal Metadata & Report Data #

Your XML structure has a Terminal node with attributes (isn, imo, name) and a nested Report node with the actual data points. We'll extract both and combine them:

# Get all Terminal nodes (handles multiple Terminals if present)
terminal_nodes <- xml_find_all(xml_doc, ".//Terminal")

# Process each Terminal into a row (or rows if multiple Reports per Terminal)
df <- map_dfr(terminal_nodes, function(terminal) {
  # Extract Terminal attributes
  terminal_meta <- tibble(
    isn = xml_attr(terminal, "isn"),
    imo = xml_attr(terminal, "imo"),
    vessel_name = xml_attr(terminal, "name")
  )
  
  # Get all Report nodes inside this Terminal
  report_nodes <- xml_find_all(terminal, ".//Report")
  
  # Extract Report data
  report_data <- map_dfr(report_nodes, function(report) {
    tibble(
      datetime = xml_text(xml_find_first(report, ".//DateTime")),
      lat = as.numeric(xml_text(xml_find_first(report, ".//Lat"))),
      lon = as.numeric(xml_text(xml_find_first(report, ".//Lon"))),
      cog = as.integer(xml_text(xml_find_first(report, ".//Cog"))),
      sog = as.integer(xml_text(xml_find_first(report, ".//Sog"))),
      voltage = as.integer(xml_text(xml_find_first(report, ".//Voltage"))),
      status = xml_text(xml_find_first(report, ".//Status"))
      # Add more fields here if your Report has additional nodes
    )
  })
  
  # Combine Terminal metadata with Report data
  bind_cols(terminal_meta, report_data)
})

Step 4: Clean Up Data Types (Optional but Recommended) #

Convert the datetime field to a proper POSIXct datetime object for easier analysis:

df <- df %>%
  mutate(datetime = as.POSIXct(datetime, format = "%d/%m/%Y %H:%M:%S"))

How This Works #

• xml_find_all locates all matching nodes in the XML document
• xml_attr pulls attribute values from a node (like the isn on Terminal)
• xml_text extracts the text content from a node (like the DateTime value)
• map_dfr applies a function to each node and binds the results into a single DataFrame (perfect for handling multiple Terminals or Reports)

This solution is flexible: if your XML has multiple Terminal nodes or multiple Report nodes per Terminal, it will still work and create a row for each unique Terminal-Report combination.

内容的提问来源于stack exchange，提问作者Dhiraj