数月未用R，求助基于规则提取字符串并保留原列逗号格式

Hey Alex, welcome back to R! Let's break down how to handle your string extraction task, plus address your question about modifying the original column vs creating a new one.

第一步：定义提取规则的函数 #

First, let's create a helper function that handles each of your three cases. We'll use stringr (part of the tidyverse) for clean string operations—it's intuitive and perfect for this kind of task:

# 先加载必要的包（如果没安装的话先运行 install.packages("tidyverse")）
library(tidyverse)

extract_target <- function(str) {
  case_when(
    # 匹配5位字母数字串，提取前3位
    str_detect(str, "^[A-Za-z0-9]{5}$") ~ str_sub(str, 1, 3),
    # 匹配6位字母数字串，跳过首字符提取后续3位（即第2-4位）
    str_detect(str, "^[A-Za-z0-9]{6}$") ~ str_sub(str, 2, 4),
    # 匹配4位数字串，提取前2位
    str_detect(str, "^[0-9]{4}$") ~ str_sub(str, 1, 2),
    # 对不符合规则的字符串，保留原内容（可根据需求调整）
    TRUE ~ str
  )
}

第二步：处理逗号分隔的列 #

Your column uses comma-separated values, so we need to split each entry, apply the extraction function to every element, then rejoin them with commas.

选项1：新建一列（推荐！）

This is the safer approach because it preserves your original data in case you need it later:

# 假设你的数据框叫df，目标列叫original_column
df <- df %>%
  mutate(
    extracted_column = str_split(original_column, ", ") %>%  # 拆分逗号分隔的字符串
      map_chr(~paste(map_chr(.x, extract_target), collapse = ", "))  # 对每个元素提取后重新合并
  )

Note: If your commas don't have a space after them (e.g., "ABC12,DEF345"), change ", " to "," in both str_split and paste.

选项2：覆盖原列

If you're sure you don't need the original data anymore, you can overwrite the column directly:

df <- df %>%
  mutate(
    original_column = str_split(original_column, ", ") %>%
      map_chr(~paste(map_chr(.x, extract_target), collapse = ", "))
  )

额外提示 #

• If you prefer base R (no tidyverse), you can achieve the same with strsplit() and sapply(), but the tidyverse syntax is more readable for complex string tasks.
• Test the function on a small subset of your data first to make sure it's handling all cases correctly—for example, check edge cases like strings that don't match any of your three rules.

内容的提问来源于stack exchange，提问作者Alex