R语言按unit #2将数据框文本列拆分为两列的实现方法

需求说明 #

现有2列结构的数据框，需以关键词unit #2首次出现的位置为分界拆分Text列：

• 拆分后第2列存储unit #2首次出现前的所有语句
• 拆分后第3列存储从unit #2首次出现位置开始到末尾的所有语句
  测试用示例数据如下：

report <- data.frame(
  Text = c(
    "unit #1 stopped at a stop sign on a road. unit #1 was speeding. unit #2 travelling southbound  in lane #2 of 3 lanes. unit #2 couldn't react in time  and crashed into unit #1. unit #2 was unmindful.", 
    "unit #1 stopped there. unit #1 was under influence of drug. unit #2 travelling northbound. unit #2 was not unmindful. unit #2 crashed into unit #1.", 
    "unit #1 was going straight. unit #1 was not speeding. unit #2 travelling southbound  in lane #1 of 2 lanes. unit #2 couldn't react in time and crashed into unit #1. unit #2 was driving fast."
  ), 
  id = 1:3
)

实现方法 #

方法1：基础R实现（无需安装额外包） #

通过regexpr定位关键词首次出现的字符位置，再用字符串截取函数完成拆分，匹配时设置fixed=TRUE将关键词作为纯文本识别，避免正则特殊字符导致匹配偏差：

# 定位每条文本中unit #2首次出现的起始位置
match_pos <- regexpr("unit #2", report$Text, fixed = TRUE)

# 截取前后两段文本，trimws用于去除首尾多余空格
report$text_before <- trimws(substr(report$Text, 1, match_pos - 1))
report$text_from_unit2 <- substr(report$Text, match_pos, nchar(report$Text))

方法2：stringr包实现（写法更简洁） #

如果日常使用tidyverse生态的工具，可以直接用正则捕获组快速拆分：

library(stringr)
library(dplyr)

report <- report %>%
  mutate(
    text_before = trimws(str_extract(Text, ".*?(?=unit #2)")),
    text_from_unit2 = str_extract(Text, "unit #2.*")
  )

结果说明 #

拆分后以第一行数据为例：

• text_before值为：unit #1 stopped at a stop sign on a road. unit #1 was speeding.
• text_from_unit2值为：unit #2 travelling southbound in lane #2 of 3 lanes. unit #2 couldn't react in time and crashed into unit #1. unit #2 was unmindful.
  完全符合拆分要求，其余行拆分逻辑一致。

内容的提问来源于stack exchange，提问作者S Das