R语言str_replace_all无法替换部分指定字符串是什么原因？

问题原因 #

stringr::str_replace_all 默认采用正则表达式规则进行模式匹配，你需要替换的两个疾病名称中包含(、)正则特殊字符，这类字符在正则语法中有分组的特殊语义，不会被识别为普通的括号文本，因此无法匹配到原始字符串中的对应内容，出现无报错但替换失败的情况。

解决方案 #

方案1：使用fixed()声明固定文本匹配（最便捷） #

给所有需要匹配的原字符串套上fixed()函数，告知函数将匹配内容作为普通固定文本处理，不需要解析正则语义，修改后的代码如下：

codelists <- data.frame(Disease = sample(c("Peripheral neuropathies (excluding cranial nerve and carpal tunnel syndromes)", "Primary Malignancy_Brain, Other CNS and Intracranial", "Venous thromboembolic disease (Excl PE)"), 15, replace = T), Codes = 1:15)
codelists <- codelists[order(codelists$Disease), ]

library(stringr)
codelists$Disease2 <- str_replace_all(codelists$Disease, c(
  fixed("Peripheral neuropathies (excluding cranial nerve and carpal tunnel syndromes)") = "Non-diabetic peripheral neuropathies (excluding cranial nerves and carpal tunnel syndrome)",
  fixed("Primary Malignancy_Brain, Other CNS and Intracranial") = "Primary malignancy brain, other CNS and intracranial",
  fixed("Venous thromboembolic disease (Excl PE)") = "Venous thromboembolism"
))

方案2：转义正则特殊字符 #

如果需要保留正则匹配能力，给所有正则特殊字符前加双反斜杠（R语言中转义符需要写两次）进行转义，示例如下：

codelists$Disease2 <- str_replace_all(codelists$Disease, c(
  "Peripheral neuropathies \\(excluding cranial nerve and carpal tunnel syndromes\\)" = "Non-diabetic peripheral neuropathies (excluding cranial nerves and carpal tunnel syndrome)",
  "Primary Malignancy_Brain, Other CNS and Intracranial" = "Primary malignancy brain, other CNS and intracranial",
  "Venous thromboembolic disease \\(Excl PE\\)" = "Venous thromboembolism"
))

方案3：整值映射替换（更适合全匹配场景） #

如果你的需求是对整个疾病名称做完全匹配替换，不需要模糊匹配，直接用命名向量映射的方式更稳妥，完全规避正则语法的坑，代码如下：

replace_map <- c(
  "Peripheral neuropathies (excluding cranial nerve and carpal tunnel syndromes)" = "Non-diabetic peripheral neuropathies (excluding cranial nerves and carpal tunnel syndrome)",
  "Primary Malignancy_Brain, Other CNS and Intracranial" = "Primary malignancy brain, other CNS and intracranial",
  "Venous thromboembolic disease (Excl PE)" = "Venous thromboembolism"
)
codelists$Disease2 <- replace_map[codelists$Disease]

内容的提问来源于stack exchange，提问作者Nao