Parallel mutate across a dynamic amount of columns in R Tidyverse-CodePudding

I have a code that will generate a tibble with

1 identifier col
1 col T with numeric values
X cols with numeric values, named A_1, A_2, A_3, ..., A_X

My output have to be

X cols named B_1, B_2, B_3, ..., B_X with numeric values such that

B_X[z] = A_X[z] * T[z]

If X = 1, the code would be straightforward:

Tibble %>%
mutate(B = A * T)

The issue is that I don't know the value of X. Basically I would need to for (i in 1:X) {Tibble %>% add_column() -> Tibble} but I am almost sure there are fancier solutions, especially because I'd like to avoid assignments and just proceed to summarize the B-cols.

Thank you.

CodePudding user response：

You could do:

df %>% 
   mutate(across(starts_with('A'), .names = "B_{str_extract(col, '[0-9] $')}") * T)

or even:

df %>% 
   mutate(across(starts_with('A'), ~.x*T, .names = "B_{str_extract(col, '[0-9] $')}"))

CodePudding user response：

You may get the columns to multiply using grep and multiplying them with T column should be straightforward. Using base R, you can do the following -

df <- data.frame(ID = 1:5, T = c(2, 4, 5, 6, 1), A_1 = c(3, 4, 5, 1, 2), 
                 A_2 = 1:5, A_3 = c(7, 8, 3, 2, 9))

cols <- grep('A_\\d ', names(df), value = TRUE)
df[sub('A', 'B', cols)] <- df$T * df[cols]
df

#  ID T A_1 A_2 A_3 B_1 B_2 B_3
#1  1 2   3   1   7   6   2  14
#2  2 4   4   2   8  16   8  32
#3  3 5   5   3   3  25  15  15
#4  4 6   1   4   2   6  24  12
#5  5 1   2   5   9   2   5   9