select value from a random column in dplyr-CodePudding

I have a tibble/dataframe df with six columns (a,b,c,d,e,f)

I randomly choose one of these six letters and assign it to an extra column:

df %>% mutate(ran=sample(c("a","b","c","d","e","f"),n(), replace=T))

So for each row I have a randomly picked column name. How can I get a value of the corresponding randomly picked column and put it to a new column?

CodePudding user response：

Here is a tidyverse solution relying on tidyr::gather.

First I create a sample tibble.

library(tidyverse)
set.seed(123)

df <-
  tibble(a = 1:5, b = a   10, c = b   10, d = c   10, e = d   10, f = e   10) %>%
  mutate(
    ran=sample(c("a","b","c","d","e","f"),n(), replace=T),
    row = 1:n()
  )

Then I pivot this dataframe and then use filter to select the right value.

df %>%
  gather(val, col, -ran, -row) %>%
  filter(ran==val) %>%
  select(row, col) %>%
  left_join(df, ., by = "row")

# A tibble: 5 x 9
      a     b     c     d     e     f ran     row   col
  <int> <dbl> <dbl> <dbl> <dbl> <dbl> <chr> <int> <dbl>
1     1    11    21    31    41    51 c         1    21
2     2    12    22    32    42    52 f         2    52
3     3    13    23    33    43    53 c         3    23
4     4    14    24    34    44    54 b         4    14
5     5    15    25    35    45    55 b         5    15