Subtract the row maximum from another column in R-CodePudding

I want to subtract the maximum value in each row by another column, and then make a new column with these values. So if I have this, and I want to subtract the max value from col1:

col1   col2    col3     col4      col5
1       3       2        3          5
3       7       4        1          2
2       2       6        8          3

To get:

col1   col2    col3     col4      col5     new
1       3       2        3          5       -4
9       7       4        1          2        2
2       2       6        8          3       -6

I'm thinking I need to use mutate and then maybe the maxstat package. Any help is appreciated!

CodePudding user response：

We could use pmax to get the max of each row and then subtract

df1$new <- df1$col1 - do.call(pmax, c(df1[-1], na.rm = TRUE))

-output

> df1
  col1 col2 col3 col4 col5 new
1    1    3    2    3    5  -4
2    9    7    4    1    2   2
3    2    2    6    8    3  -6

Or with tidyverse

library(dplyr)
library(purrr)
df1 %>%
   mutate(new = col1 - exec(pmax, !!! rlang::syms(names(.)[-1]),
         na.rm = TRUE))
  col1 col2 col3 col4 col5 new
1    1    3    2    3    5  -4
2    9    7    4    1    2   2
3    2    2    6    8    3  -6

Or we could use

df1 %>% 
  mutate(new = col1 - invoke(pmax, across(-col1), na.rm = TRUE))
  col1 col2 col3 col4 col5 new
1    1    3    2    3    5  -4
2    9    7    4    1    2   2
3    2    2    6    8    3  -6

data

df1 <- structure(list(col1 = c(1, 9, 2), col2 = c(3L, 7L, 2L), col3 = c(2L, 
4L, 6L), col4 = c(3L, 1L, 8L), col5 = c(5L, 2L, 3L)),
 row.names = c(NA, 
-3L), class = "data.frame")