R语言编写正则表达式移除数据框行值中[preliminary_arm_1]的方法

你需要删除的是完全固定的字符串[preliminary_arm_1]，不需要编写复杂正则，优先选固定字符串匹配的写法，能完全避开方括号这类正则特殊字符的转义坑，稳定性和运行效率都更高。

推荐实现（零正则风险） #

用R基础自带的gsub()做全局替换，开启fixed = TRUE参数，让函数按字面意思匹配目标字符串，不解析正则规则：

# 测试数据复现
branching_loc <- c("([preliminary_arm_1][antibiotic_arm] = '1') and [was_review_done]='1'",
                   "[preliminary_arm_1][antibiotic_arm]  = '1' and [was_review_done]=='1'",
                   "[preliminary_arm_1][antibiotic_arm]  = '1' and [was_review_done]=='1'",
                   "[preliminary_arm_1][antibiotic_arm]  = '1' and [was_review_done]=='1'")
df <- data.frame(branching_loc)

# 批量移除目标片段
df$branching_loc <- gsub(
  pattern = "[preliminary_arm_1]",
  replacement = "",
  x = df$branching_loc,
  fixed = TRUE
)

处理后列的输出结果如下，其余所有语法、字段标记、符号都完全保留，仅删掉了目标片段：

> df$branching_loc
[1] "([antibiotic_arm] = '1') and [was_review_done]='1'"
[2] "[antibiotic_arm]  = '1' and [was_review_done]=='1'"
[3] "[antibiotic_arm]  = '1' and [was_review_done]=='1'"
[4] "[antibiotic_arm]  = '1' and [was_review_done]=='1'"

正则写法（按需选用） #

如果业务场景要求必须用正则实现，注意[和]是正则的特殊元字符，需要用双反斜杠转义才能被识别为普通文本，写法如下：

df$branching_loc <- gsub(
  pattern = "\\[preliminary_arm_1\\]",
  replacement = "",
  x = df$branching_loc
)

运行结果和上面固定匹配的写法完全一致。

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