在R中过滤DataFrame：保留warning为HAS且messages列非NA的行

问题背景 #

现有如下定义的tibble数据框df：

warning    = c("HAS","NO","HAS","HAS","HAS")
validation = c("OK","OK","WARNING","WARNING","WARNING") 
a_b_c_messages1 = c(NA,NA,"good catch",NA,NA)
D_E_f_messages2 = c(NA,NA,NA,"NOT BAD",NA)
g_h_I_messages3 = c(NA,NA,NA,NA,"BETTER")
j_k_l_messages4 = c(NA,NA,NA,NA,NA)

df = tibble(warning,validation,a_b_c_messages1,
            D_E_f_messages2,g_h_I_messages3,
            j_k_l_messages4)

数据结构如下：

# A tibble: 5 × 6
  warning validation a_b_c_messages1 D_E_f_messages2 g_h_I_messages3 j_k_l_messages4
  <chr>   <chr>      <chr>           <chr>           <chr>           <lgl>          
1 HAS     OK         NA              NA              NA              NA              
2 NO      OK         NA              NA              NA              NA              
3 HAS     WARNING    good catch      NA              NA              NA              
4 HAS     WARNING    NA              NOT BAD         NA              NA              
5 HAS     WARNING    NA              NA              BETTER          NA  

需要筛选出满足以下两个条件的行：

1. warning列值为"HAS"
2. 任意列名包含"messages"的列的值不为NA

解决方案 #

使用dplyr包的filter结合if_any函数可以高效实现需求：

library(dplyr)

filtered_df <- df %>%
  filter(
    warning == "HAS",
    if_any(contains("messages"), ~ !is.na(.x))
  )

代码解释 #

• warning == "HAS"：直接筛选出warning列等于"HAS"的行
• contains("messages")：自动匹配所有列名包含"messages"的列，无需手动指定列名，适配列名变化
• if_any(..., ~ !is.na(.x))：检查匹配到的列中是否存在任意一列的值不为NA，满足则保留该行

结果展示 #

运行上述代码后得到的filtered_df如下：

# A tibble: 3 × 6
  warning validation a_b_c_messages1 D_E_f_messages2 g_h_I_messages3 j_k_l_messages4
  <chr>   <chr>      <chr>           <chr>           <chr>           <lgl>          
1 HAS     WARNING    good catch      NA              NA              NA              
2 HAS     WARNING    NA              NOT BAD         NA              NA              
3 HAS     WARNING    NA              NA              BETTER          NA  

内容的提问来源于stack exchange，提问作者Homer Jay Simpson