在R的dplyr中检测字符串后生成含修正值的新列

问题描述 #

需要在R中通过mutate检测某列（示例为Name列）的字符串，将部分匹配或完全匹配的内容替换为修正值，但要保留原列作为参考，生成仅包含修正后值的新列。原代码直接修改了Name列，错误尝试则会把整个数据框赋值给新列，导致类型不匹配的问题。

原修改原列的代码 #

df <- data.frame(Name = c("Jim","Bob","Sue","Sally","Jimmm","Boob","Suezi","Sallyyyy","Jim","Bob","Sue","Sally"),
Period = c("P1","P1","P1","P1","P2","P2","P2","P2","P3","P3","P3","P3"),
Value = c(150, 200, 325, 120, 760,245,46,244,200, 325, 120, 760))

df <- df %>% 
  mutate(Name = case_when(
  str_detect(Name, "Jim") ~ "Jim",
str_detect(Name, "Sue") ~ "Sue",
  TRUE ~ Name)) %>%
  mutate(across(Name, str_replace, "Sallyyyy", "Sally")) 

错误尝试代码 #

df$test <- df %>% 
  mutate(Name = case_when(
  str_detect(Name, "Jim") ~ "Jim",
  TRUE ~ Name)) %>%
  mutate(across(Name, str_replace, "Sallyyyy", "Sally")) 

解决方案 #

核心是在mutate中直接定义新列名，基于原列执行修正逻辑，而非修改原列。以下是两种可行的实现方式：

方式一：合并修正逻辑到单个mutate #

把所有字符串匹配规则整合到同一个case_when中，一步生成修正后的新列，代码更简洁：

# 加载依赖包
library(dplyr)
library(stringr)

df <- data.frame(Name = c("Jim","Bob","Sue","Sally","Jimmm","Boob","Suezi","Sallyyyy","Jim","Bob","Sue","Sally"),
Period = c("P1","P1","P1","P1","P2","P2","P2","P2","P3","P3","P3","P3"),
Value = c(150, 200, 325, 120, 760,245,46,244,200, 325, 120, 760))

# 生成新列Name_clean，保留原Name列
df <- df %>% 
  mutate(
    Name_clean = case_when(
      str_detect(Name, "Jim") ~ "Jim",
      str_detect(Name, "Sue") ~ "Sue",
      str_detect(Name, "Sally") ~ "Sally",
      TRUE ~ Name
    )
  )

方式二：分两步修正新列 #

如果需要保留原有的分步修正逻辑，可以先创建初步修正的新列，再对该新列做二次处理：

df <- df %>% 
  mutate(Name_clean = case_when(
    str_detect(Name, "Jim") ~ "Jim",
    str_detect(Name, "Sue") ~ "Sue",
    TRUE ~ Name
  )) %>%
  mutate(Name_clean = str_replace(Name_clean, "Sallyyyy", "Sally"))

关键说明 #

• 直接在mutate中指定新列名（如Name_clean），所有操作基于原Name列执行，不会修改原列内容。
• 错误尝试中df$test <- ...会把整个处理后的data.frame赋值给单列，导致类型不匹配；正确方式是在管道流中直接添加新列，再重新赋值给原数据框。

内容的提问来源于stack exchange，提问作者John Conor