Calculating values in a column as a function of previous column value in same row position

user_10794317 0
Viewed 69

I have many columns for which i want to calculate value based on function (1 + x) ^ k, where x is a value from a particular column and k is index of column for which we try to calculate. I want to calculate this only for a subset of all the columns of the table.

For axample:

df = data.frame(A = c(0.1, 0.05, 0.2), B = c(1, 1, 1), C = c(NA, NA, NA), D = c(NA, NA, NA)

I want to apply the function only to columns C and D, using values from column A.

For example df[1,4] would be calculated as (1 + (-0.1)^4 because 4 is index of column D.

An alternative interpretation is that for the selected columns (C and D in this example), the value is value from previous column multiplied by (1+x), i.e. df[1,4] = df[1,3] * (1 + (-0.1)), giving the same result

dataframe r function
4 Answers

One dplyr option could be:

df %>%
 rowwise() %>%
 mutate((1 + A)^(across(C:D, ~ replace(., is.na(.), 1)) * which(names(.) %in% c("C", "D"))))

      A     B     C     D
  <dbl> <dbl> <dbl> <dbl>
1  0.1      1  1.33  1.46
2  0.05     1  1.16  1.22
3  0.2      1  1.73  2.07

Or if the C and D columns are always NAs:

df %>%
 rowwise() %>%
 mutate((1 + A)^(1^across(C:D) * which(names(.) %in% c("C", "D"))))
user_59645570

Maybe something like this:

f = function(df, target_cols = c("C", "D"), index_col = "A"){
  
  # df is your data frame 
  # target_cols is a vector of columns you want apply the function (1 + x)^k 
  # index_col is the the x in (1 + x)^k

  stopifnot(all(target_cols %in% names(df)))
  stopifnot(index_col %in% names(df))
  
  # Get index of target columns i.e k
  which_col = which(names(df) %in% target_cols)
  
  # Loop over columns
  for(i in which_col){
    df[, i] = (1 + df[, index_col])^i
  }
  
  return(df)
  
}

df = data.frame(A = c(0.1, 0.05, 0.2), B = c(1, 1, 1), C = c(NA, NA, NA), D = c(NA, NA, NA))
f(df)                

     A B        C        D
1 0.10 1 1.331000 1.464100
2 0.05 1 1.157625 1.215506
3 0.20 1 1.728000 2.073600
user_87415750

The function you described is defined as follows:

func1 = function(df, i) {
  (1+df[1]^i)
}
user_140874780

An option in base R would be

df[3:4] <- (1 + df$A)^col(df)[, 3:4]

Or with Reduce

df[c('C', 'D')] <- lapply(match(c('C', 'D'), names(df)), function(i) 
     Reduce(function(x, y) (1 + y)^i, df[, 'A'],
          accumulate = TRUE, init = df$A[1])[-1])

df
#     A B        C        D
#1 0.10 1 1.331000 1.464100
#2 0.05 1 1.157625 1.215506
#3 0.20 1 1.728000 2.073600

Or using map/accumulate

library(purrr)
library(dplyr)
map_dfc(set_names(match(c('C', 'D'), names(df)), names(df)[3:4]), ~ {
            i <- .x
            accumulate(df$A,  ~(1 + .y)^i, 
              .init = first(df$A))[-1]}) %>% 
   bind_cols(df[1:2], .)
#     A B        C        D
#1 0.10 1 1.331000 1.464100
#2 0.05 1 1.157625 1.215506
#3 0.20 1 1.728000 2.073600
user_37322710
Related