R语言：分组处理字符串，移除长串中非目标匹配子串

需求说明 #

现有一个R语言数据框，其中V1列为长字符串，每行包含至少两个可匹配的子串；V2列为该行需要保留的目标子串。要求：对每组相同V1值的行，将该行V1中除对应V2之外的同组其他V2子串移除，仅保留包含目标V2的长串部分。

示例输入 #

library(qdap)
library(magrittr)
t <- read.table(text="
V1,V2
  Video of all presentations and discussionPart 1Part 2Video,Part 1
  Video of all presentations and discussionPart 1Part 2Video,Part 2
  Video of all presentations and discussionPart 1Part 2Video,Video
  Background - PDFVideo Soil management and update (pdf)Video,PDF
  Background - PDFVideo Soil management and update (pdf)Video,Video
  Background - PDFVideo Soil management and update (pdf)Video,Soil management and update (pdf)
  Background - PDFVideo Soil management and update (pdf)Video,Video
", header=T, sep = ",")

尝试的无效代码 #

t%>%  
  split(.,.$V1)%>% 
    lapply(.,function(x){(unique(x$V2))})%>%
      lapply(.,function(y){mgsub(pattern=y[[1]],replacement="",names(y))})

期望输出 #

t <- read.table(text="
V1,V2
Video of all presentations and discussionPart 1,Part 1
Video of all presentations and discussionPart 2,Part 2
Video of all presentations and discussionVideo,Video
Background - PDF,PDF
Background - Video,Video
Background - Soil management and update (pdf),Soil management and update (pdf)
Background - Video,Video
", header=T, sep = ",")

解决方案 #

使用dplyr分组处理，结合stringr的字符串操作实现需求：

library(dplyr)
library(stringr)

# 读取示例数据（若已加载可跳过）
t <- read.table(text="
V1,V2
  Video of all presentations and discussionPart 1Part 2Video,Part 1
  Video of all presentations and discussionPart 1Part 2Video,Part 2
  Video of all presentations and discussionPart 1Part 2Video,Video
  Background - PDFVideo Soil management and update (pdf)Video,PDF
  Background - PDFVideo Soil management and update (pdf)Video,Video
  Background - PDFVideo Soil management and update (pdf)Video,Soil management and update (pdf)
  Background - PDFVideo Soil management and update (pdf)Video,Video
", header=T, sep=",", stringsAsFactors=F)

# 核心处理逻辑
result <- t %>%
  group_by(V1) %>%
  mutate(
    # 生成当前行需要移除的子串正则：组内唯一V2排除当前行V2
    remove_pattern = str_c(setdiff(unique(V2), V2), collapse = "|"),
    # 移除匹配子串并清理多余空格
    new_V1 = str_remove_all(V1, remove_pattern) %>% str_squish()
  ) %>%
  ungroup() %>%
  # 替换原V1列并整理输出
  select(new_V1, V2) %>%
  rename(V1 = new_V1)

# 输出结果
print(result)

代码说明 #

1. 分组：按V1分组，确保仅处理同组内的子串；
2. 生成移除规则：用setdiff筛选出当前组内除当前行V2外的所有唯一子串，拼接成正则表达式；
3. 移除子串：用str_remove_all批量移除匹配的子串，str_squish清理可能出现的多余空格；
4. 整理输出：替换原V1列并取消分组，得到符合要求的结果。

内容的提问来源于stack exchange，提问作者CrunchyTopping