r - 有没有办法部分匹配文本/字符串并返回 R 中的完整值

Question

我有一个数据框，我需要根据另一列中匹配的值填充一些列。

运行此代码以获取示例数据集

sample_data <- structure(list(Temp = c(NA_character_, NA_character_, NA_character_, 
NA_character_, NA_character_), Wind = c(NA_character_, NA_character_, 
NA_character_, NA_character_, NA_character_), NodeID = c(3, 5, 
6, 8, 9), node_path = c("Temp <= 82 , Wind <= 6.9", "Temp <= 82 , Wind > 6.9 , Temp <= 77", 
"Temp <= 82 , Wind > 6.9 , Temp > 77", "Temp > 82 , Wind <= 10.3", 
"Temp > 82 , Wind > 10.3")), row.names = c(NA, -5L), class = c("tbl_df", 
"tbl", "data.frame"))

这就是我想要实现的目标。沿 node_path 匹配列名称 Temp 和 Wind，并将特定的匹配值返回到这些列。我尝试使用 str_extract_all(node_path, pattern = "^Temp.*")，但这会返回 node_path 下的整个单元格值。知道如何实现这一目标吗？

Answer 1

一个选项(无需 reshape )是通过在“Temp”、“Wind”NA 之间循环来根据列名称 (cur_column()) 提取元素列并通过循环 list 来粘贴 (toString) str_extract_all 中的 list 元素> 带有 map

library(dplyr)
library(purrr)
library(stringr)
sample_data %>% 
    mutate(across(c(Temp, Wind), ~ map_chr(str_extract_all(node_path, 
         str_c(cur_column(), "\\D+[0-9.]+")), toString)))

-输出

# A tibble: 5 x 4
  Temp                   Wind         NodeID node_path                           
  <chr>                  <chr>         <dbl> <chr>                               
1 Temp <= 82             Wind <= 6.9       3 Temp <= 82 , Wind <= 6.9            
2 Temp <= 82, Temp <= 77 Wind > 6.9        5 Temp <= 82 , Wind > 6.9 , Temp <= 77
3 Temp <= 82, Temp > 77  Wind > 6.9        6 Temp <= 82 , Wind > 6.9 , Temp > 77 
4 Temp > 82              Wind <= 10.3      8 Temp > 82 , Wind <= 10.3            
5 Temp > 82              Wind > 10.3       9 Temp > 82 , Wind > 10.3                    

---

或者可以在 base R 中使用 regmatches/regexpr 来根据模式提取值，然后粘贴 toString

sample_data[1:2] <- lapply(names(sample_data)[1:2], function(x) 
      sapply(regmatches(sample_data$node_path, gregexpr(paste0(x, 
    "\\D+[0-9.]+"), sample_data$node_path)), toString))

---

或者另一种选择是使用 separate_rows 拆分行，然后执行 pivot_wider 重新整形回宽格式

library(tidyr)
sample_data %>% 
    select(node_path, NodeID) %>%
    separate_rows(node_path, sep="\\s*,\\s*") %>% 
    mutate(colnm = word(node_path, 1)) %>%
    group_by(colnm, NodeID) %>% 
    summarise(new = str_c(node_path, collapse=", "), .groups = 'drop') %>%  
    pivot_wider(names_from = colnm, values_from = new) %>%    
    left_join(sample_data %>% 
                 select(NodeID, node_path))

-输出

# A tibble: 5 x 4
  NodeID Temp                   Wind         node_path                           
   <dbl> <chr>                  <chr>        <chr>                               
1      3 Temp <= 82             Wind <= 6.9  Temp <= 82 , Wind <= 6.9            
2      5 Temp <= 82, Temp <= 77 Wind > 6.9   Temp <= 82 , Wind > 6.9 , Temp <= 77
3      6 Temp <= 82, Temp > 77  Wind > 6.9   Temp <= 82 , Wind > 6.9 , Temp > 77 
4      8 Temp > 82              Wind <= 10.3 Temp > 82 , Wind <= 10.3            
5      9 Temp > 82              Wind > 10.3  Temp > 82 , Wind > 10.3