I have an extended dataset that I would simplify like this:

BEZ <- c(0.5, 1.5)
var <- c(0, 1.5 )
bar <- c(3, 1.5)

BEZ1 <- c(0, 0.5)
var1 <- c(4, 4)
bar1 <- c(4, 4.5)

dat <- data.frame(BEZ, var, bar, BEZ1, var1, bar1)
dat

What I would like to do is to add two final columns reporting two sums:

1. the sum of every middle columns different from those BEZ;
2. the division between the first result and the sum of columns having 'BEZ' in their names.

I have used this way:

scores = dat %>% 
  select(-starts_with('BEZ')) %>% 
  #replace(is.na(.), 0) %>% 
  mutate(score_1 = rowSums(.), 
         score_2 = score_1/rowSums(dat %>% select(starts_with('BEZ'))))

new = cbind(dat, scores[, 5:6])
new

But I am looking for a way that would be easier and avoiding to create different chunks of code. Could you suggest any other different alternatives?

Thanks

CodePudding user response：

You can use across() to select columns on which you want to do row-summing:

library(dplyr)

dat %>%
  mutate(score_1 = rowSums(across(!starts_with('BEZ'))),
         score_2 = score_1 / rowSums(across(starts_with('BEZ'))))

#   BEZ var bar BEZ1 var1 bar1 score_1 score_2
# 1 0.5 0.0 3.0  0.0    4  4.0    11.0   22.00
# 2 1.5 1.5 1.5  0.5    4  4.5    11.5    5.75