R处理XML提取结局指标时缺失节点赋值NA问题求助

问题根源 #

你之前的代码应该是批量提取所有<measure>、<time_frame>、<description>节点的文本后直接合并成DataFrame。但因为部分<secondary_outcome>没有<description>节点，提取到的描述数量远少于结局总数，R会自动循环补齐缺失值，导致描述和对应结局错位。

解决方案 #

核心思路是以单个结局节点为单位提取字段：遍历每个<primary_outcome>和<secondary_outcome>节点，单独提取该节点下的三个字段，没有对应子节点时直接赋值NA。这样每个结局的三个字段能严格对应，不会错位。

用xml2包实现（推荐，语法更简洁） #

# 加载所需包
library(xml2)
library(dplyr)
library(writexl)

# 读取XML文件
doc <- read_xml("你的XML文件路径.xml")

# 定义函数：从单个结局节点提取信息
extract_outcome <- function(node) {
  # 提取当前节点下的子节点文本，找不到则返回空字符串，再转为NA
  measure <- xml_text(xml_find_first(node, "./measure"), trim = TRUE)
  time_frame <- xml_text(xml_find_first(node, "./time_frame"), trim = TRUE)
  description <- xml_text(xml_find_first(node, "./description"), trim = TRUE)
  
  tibble(
    measure = ifelse(measure == "", NA_character_, measure),
    time_frame = ifelse(time_frame == "", NA_character_, time_frame),
    description = ifelse(description == "", NA_character_, description)
  )
}

# 提取主要结局并标记类型
primary_df <- xml_find_all(doc, "//primary_outcome") %>%
  purrr::map_dfr(extract_outcome) %>%
  mutate(outcome_type = "primary")

# 提取次要结局并标记类型
secondary_df <- xml_find_all(doc, "//secondary_outcome") %>%
  purrr::map_dfr(extract_outcome) %>%
  mutate(outcome_type = "secondary")

# 合并所有数据
final_df <- bind_rows(primary_df, secondary_df)

# 导出为Excel文件
write_xlsx(final_df, "结局提取结果.xlsx")

关键代码解释 #

• xml_find_first(node, "./measure")：针对单个结局节点，只查找它下面的<measure>子节点，找不到返回NA
• map_dfr：遍历所有结局节点，把每个节点的提取结果合并成一个DataFrame
• ifelse(measure == "", NA_character_, measure)：把空字符串转为标准的NA，方便后续处理

用旧版XML包实现 #

如果习惯用XML包，逻辑完全一致：

library(XML)
library(dplyr)
library(writexl)

# 自定义空值替换函数
`%||%` <- function(a, b) if (is.null(a)) b else a

# 读取XML文件
doc <- xmlParse("你的XML文件路径.xml")

# 定义提取函数
extract_outcome <- function(node) {
  measure <- xmlValue(getNodeSet(node, "./measure")[[1]], trim = TRUE) %||% NA_character_
  time_frame <- xmlValue(getNodeSet(node, "./time_frame")[[1]], trim = TRUE) %||% NA_character_
  description <- xmlValue(getNodeSet(node, "./description")[[1]], trim = TRUE) %||% NA_character_
  
  data.frame(measure, time_frame, description, stringsAsFactors = FALSE)
}

# 提取主要结局
primary_nodes <- getNodeSet(doc, "//primary_outcome")
primary_df <- bind_rows(lapply(primary_nodes, extract_outcome)) %>%
  mutate(outcome_type = "primary")

# 提取次要结局
secondary_nodes <- getNodeSet(doc, "//secondary_outcome")
secondary_df <- bind_rows(lapply(secondary_nodes, extract_outcome)) %>%
  mutate(outcome_type = "secondary")

# 合并导出
final_df <- bind_rows(primary_df, secondary_df)
write_xlsx(final_df, "结局提取结果.xlsx")

内容的提问来源于stack exchange，提问作者celilati