My problem is pretty similar to the following post
In my database i have multiple columns containing values like M0, M6, M12… and so on. And i have columns having those names M0, M6, M12…
I would like to replace the first columns containing the M0… with the value corresponding in the column
As an understanding example i have :
df <- data.frame(time=c("M0","M12","M0","M0","M12"),
M0=c(1,2,4,5,1),
M12=c(8,5,9,2,1))
df
time M0 M12
1 M0 1 8
2 M12 2 5
3 M0 4 9
4 M0 5 2
5 M12 1 1
As mentioned i found a post with interesting answers, and one in particular with the use of cur_data()
df2 <- df %>%
rowwise %>%
mutate(newd = cur_data()[[cur_data()$time]]) %>%
ungroup
df2
# A tibble: 5 × 4
time M0 M12 newd
<chr> <dbl> <dbl> <dbl>
1 M0 1 8 1
2 M12 2 5 5
3 M0 4 9 4
4 M0 5 2 5
5 M12 1 1 1
It works, but takes a really long time and also apparently the cur_data is deprecated in favor of pick() which i don’t get how to replace to make it work
Any advice on this would be much appreciated !
>Solution :
rowwise is a pretty slow operation in dplyr, even for moderately sized data frames. One possibility is to iterate in parallel over time and the row number and simply select the value:
library(dplyr)
library(purrr)
df |>
mutate(newd = map2_dbl(row_number(), time, \(idx, var) df[idx, var]))