如何将列值转换为数据表中的新列？

Question

I have a table that looks like this:

> 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

I want to take the values of is_NL and make them into separate columns (eg is_NL_0 , is_NL_1 , is_NL_2 ), and, for now, fill them with the values from counts and nrows semi-colon separated (eg 32968;685 ). I've been using tidyr 's pivot_wider to do this but, because I'm inexperienced with this package, I've been having a little trouble:

> 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.

I'm going to keep working on this but would like to know how I could do this in a way that would make sense.

Answer 1

Not familiar with tidyr but you could do:

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

Data

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

This should work fine for you case.

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