R 逐行连接 data.table 中各列的名称和值

发布于04月02日

类似于this个问题，我想连接列值的子集.但是，我还想用列的名称作 for each 值的前缀，用":"分隔名称和值，并控制用于分隔名称/值对的字符.

例如，这给了我值的串联:

library(data.table)
d <- data.table(x = c(1, 2), y = c(3, 4))

labs_to_get <- c("x", "y")


d[
  ,
  .(x, y, labs = do.call(paste, c(.SD, sep = ", "))),
  .SDcols = labs_to_get
]
#>    x y labs
#> 1: 1 3 1, 3
#> 2: 2 4 2, 4

但我在找这个:

d[
  ,
  .(x, y, labs = c("x: 1, y: 3", "x: 2, y: 4"))
]
#>    x y       labs
#> 1: 1 3 x: 1, y: 3
#> 2: 2 4 x: 2, y: 4

^{由reprex package(v2.0.1)于2022-04-01创建}

推荐答案

我们可以用Map个名字对应paste个名字

d[,
  .(x, y, labs = do.call(paste, c(Map(function(u, v) 
       paste0(v, ": ", u), .SD, labs_to_get), sep = ", "))),
  .SDcols = labs_to_get
]

-输出

     x     y       labs
   <num> <num>     <char>
1:     1     3 x: 1, y: 3
2:     2     4 x: 2, y: 4

或者另一个选项是write.dcf

d[, labs := do.call(paste, 
        c(as.list(setdiff(capture.output(write.dcf(.SD)), "")), 
     sep = ", ")), 1:nrow(d)]
> d
       x     y       labs
   <num> <num>     <char>
1:     1     3 x: 1, y: 3
2:     2     4 x: 2, y: 4

或者使用apply在行上循环

d[, labs := apply(.SD, 1, \(x) paste(names(x), x, sep = ": ", 
    collapse = ", ")), .SDcols = labs_to_get]

或者用tidyverse

library(dplyr)
library(purrr)
library(stringr)
d %>% 
  mutate(labs = invoke(str_c, c(across(all_of(labs_to_get),  
      ~str_c(cur_column(), ": ", .x)), sep = ", ")))
       x     y       labs
   <num> <num>     <char>
1:     1     3 x: 1, y: 3
2:     2     4 x: 2, y: 4