如何将列值转换为数据表中的新列？

Question

我有一张看起来像这样的表：

> dt
                 variant_id           transcript_id is_NL counts nrows
  1: chr10_60842447_A_G_b38 chr10_60871326_60871443     0  32968   685
  2: chr10_60842447_A_G_b38 chr10_60871326_60871443     1   1440    20
  3: chr10_60842447_A_G_b38 chr10_60871326_60871443     2    337     1
  4: chr10_60846892_G_A_b38 chr10_60871326_60871443     0  33157   690
  5: chr10_60846892_G_A_b38 chr10_60871326_60871443     1   1251    15
 ---                                                                  
227:  chr5_96832353_G_T_b38  chr5_96727531_96729611     1  33504   572
228:  chr5_96832353_G_T_b38  chr5_96727531_96729611     2   3352    52
229:  chr5_96834213_T_G_b38  chr5_96727531_96729611     0 110144  2208
230:  chr5_96834213_T_G_b38  chr5_96727531_96729611     1  33252   564
231:  chr5_96834213_T_G_b38  chr5_96727531_96729611     2   3352    52

我想取is_NL的值并将它们放入单独的列（例如is_NL_0 、 is_NL_1 、 is_NL_2 ），现在，用counts和nrows分号分隔的值填充它们（例如32968;685 ）。 我一直在使用tidyr的pivot_wider来做到这一点，但是，因为我对这个包没有经验，所以我遇到了一些麻烦：

> dt %>% pivot_wider(-c(transcript_id, variant_id), names_from = "is_NL", values_from = paste0(dt$counts, ";", dt$nrows), names_prefix = "NL_") %>% as.data.table
Error: Unknown columns `32968;685`, `1440;20`, `337;1`, `33157;690`, `1251;15` and ... 
Run `rlang::last_error()` to see where the error occurred.

我将继续致力于此，但想知道我如何以一种有意义的方式做到这一点。

Answer 1

不熟悉tidyr但你可以这样做：

dt[, tmp := paste(counts, nrows, sep = ";")
   ][, dcast(.SD, transcript_id + variant_id ~ is_NL, value.var = "tmp")]

             transcript_id             variant_id           0         1       2
1: chr10_60871326_60871443 chr10_60842447_A_G_b38   32968;685   1440;20   337;1
2: chr10_60871326_60871443 chr10_60846892_G_A_b38   33157;690   1251;15    <NA>
3:  chr5_96727531_96729611  chr5_96832353_G_T_b38        <NA> 33504;572 3352;52
4:  chr5_96727531_96729611  chr5_96834213_T_G_b38 110144;2208 33252;564 3352;52

数据

library(data.table)
dt <- fread("           variant_id           transcript_id is_NL counts nrows
chr10_60842447_A_G_b38 chr10_60871326_60871443     0  32968   685
chr10_60842447_A_G_b38 chr10_60871326_60871443     1   1440    20
chr10_60842447_A_G_b38 chr10_60871326_60871443     2    337     1
chr10_60846892_G_A_b38 chr10_60871326_60871443     0  33157   690
chr10_60846892_G_A_b38 chr10_60871326_60871443     1   1251    15
chr5_96832353_G_T_b38  chr5_96727531_96729611     1  33504   572
chr5_96832353_G_T_b38  chr5_96727531_96729611     2   3352    52
chr5_96834213_T_G_b38  chr5_96727531_96729611     0 110144  2208
chr5_96834213_T_G_b38  chr5_96727531_96729611     1  33252   564
chr5_96834213_T_G_b38  chr5_96727531_96729611     2   3352    52")

Answer 2

这应该适合你的情况。

library(tidyverse)

df_example <- tibble::tribble(~variant_id,~transcript_id, ~is_NL, ~counts, ~ nrows,
"chr10_60842447_A_G_b38", "chr10_60871326_60871443",     0,  32968,   685,
 "chr10_60842447_A_G_b38", "chr10_60871326_60871443",    1 ,  1440  ,  20,
 "chr10_60842447_A_G_b38" ,"chr10_60871326_60871443",     2,    337  ,   1,
 "chr10_60846892_G_A_b38" ,"chr10_60871326_60871443",     0 , 33157   ,690,
 "chr10_60846892_G_A_b38" ,"chr10_60871326_60871443",     1  , 1251    ,15)

df_example %>%
  mutate(counts = counts %>% as.character(),
         nrows = nrows %>% as.character()) %>% 
  unite("result",counts,nrows,sep = ";") %>% 
  pivot_wider(names_from = is_NL,values_from = result)


# A tibble: 2 x 5
  variant_id             transcript_id           `0`       `1`     `2`  
  <chr>                  <chr>                   <chr>     <chr>   <chr>
1 chr10_60842447_A_G_b38 chr10_60871326_60871443 32968;685 1440;20 337;1
2 chr10_60846892_G_A_b38 chr10_60871326_60871443 33157;690 1251;15 NA