基于部分匹配替换字符串——tidyverse实现报错问题

问题描述 #

有一个字符类型列wind_speed_m_s，部分观测值的字符串中包含"ph"（包括带空格的形式如"0.8 ph"），需要将所有包含"ph"的字符串替换为NA。

原始数据 #

wind_speed_m_s <- c("1.7", "0.7", "0", "0.6", "0.4", "1.2", "1.9", "1.3", "2.0, gust to 3.7", 
"0.5", "1.8", "1.4", "3.4", "2.8", "1.6", "2", NA, "0.9", "0.8", 
"1", "1.1", "2.6", "2.4", "1.1ph", "1.7 kt", "2.1", "1.5", "0ph", 
"3", ".4 /s", "0.3", "2.3", "0.2", "3.3", "3.9ph", "1.5ph", "1ph", 
"2ph", "1.7ph", "0.8 ph", "1.5 ph", "2.2", "1.9 k/hr", "2.5", 
"NA", "0.4/s", "1/s")

data_raw <- tibble(wind_speed_m_s)

错误代码及原因 #

尝试的代码：

date <- data_raw %>%
  mutate(wind_speed_m_s = str_replace(wind_speed_m_s, pattern = str_detect("ph"), "NA"))

报错信息：

Error in `mutate()`:
! Problem while computing `wind_speed_m_s = str_replace(wind_speed_m_s, pattern = str_detect("ph"), "NA")`.
Caused by error in `type()`:
! argument "pattern" is missing, with no default
Backtrace:
  1. ... %>% ...
  9. stringr::str_detect("ph")
10. stringr:::type(pattern)

错误原因：

• str_detect需要两个参数：待检测的字符串向量和匹配模式，代码中只传入了"ph"，缺少第一个参数，导致报错。
• str_replace的pattern参数需要是正则表达式字符串，而非逻辑值，即使str_detect调用正确，返回的逻辑向量也无法作为str_replace的pattern使用。

正确实现方法 #

方法一：用if_else结合str_detect #

通过逻辑判断筛选出包含"ph"的元素，替换为NA：

data_clean <- data_raw %>%
  mutate(wind_speed_m_s = if_else(str_detect(wind_speed_m_s, "ph"), NA_character_, wind_speed_m_s))

注：使用NA_character_而非普通NA，确保列的字符类型保持一致，避免类型转换问题。

方法二：用str_replace_all配合正则表达式 #

匹配所有包含"ph"的字符串，直接替换为NA：

data_clean <- data_raw %>%
  mutate(wind_speed_m_s = str_replace_all(wind_speed_m_s, ".*ph.*", NA_character_))

正则表达式.*ph.*表示匹配任意包含"ph"的字符串（.*匹配任意长度的任意字符）。

内容的提问来源于stack exchange，提问作者Eizy