Iteration through rows of a dataframe within group of columns in R

问题

I have a dataframe df with 6 fields A,B,C,D,E & F. My requirement is to create a new column G which is equal to the previous value(C) + previous value(D) + previous (G) - F. But this needs to be implemented at a group level through columns A & B (group by A & B). In case it is the first row within the group then the value in column G should be equal to E.

Sample Df -

A   B   C   D   E   F
1   2   100 200 300 0
1   2   110 210 310 10
1   2   120 130 300 10
1   1   140 150 80  0
1   1   50  60  80  20
1   1   50  60  80  20

Output -

A   B   C   D   E   F   G
1   2   100 200 300 0   300
1   2   110 210 310 10  590
1   2   120 130 300 10  900
1   1   140 150 80  0   80
1   1   50  60  80  20  350
1   1   50  60  80  20  440

Please provide a suitable solution.

回答1:

Here is one option with dplyr where we group by 'A', 'B', take the lag of 'C', 'D', 'E' add (+) them, and subtract from 'F', and coalesce with the 'E' column

library(dplyr)
df1 %>%
    group_by(A, B) %>%
     mutate(G = coalesce(lag(C) + lag(D) + lag(E) - F, E))

-output

# A tibble: 6 x 7
# Groups:   A, B [2]
#      A     B     C     D     E     F     G
#  <int> <int> <int> <int> <int> <int> <int>
#1     1     2   100   200   300     0   300
#2     1     2   110   210   310    10   590
#3     1     2   120   130   300    10   620
#4     1     1   140   150    80     0    80
#5     1     1    50    60    80    20   350
#6     1     1    50    60    80    20   170

data

df1 <- structure(list(A = c(1L, 1L, 1L, 1L, 1L, 1L), B = c(2L, 2L, 2L, 
1L, 1L, 1L), C = c(100L, 110L, 120L, 140L, 50L, 50L), D = c(200L, 
210L, 130L, 150L, 60L, 60L), E = c(300L, 310L, 300L, 80L, 80L, 
80L), F = c(0L, 10L, 10L, 0L, 20L, 20L)), class = "data.frame",
row.names = c(NA, 
-6L))

来源：https://stackoverflow.com/questions/65115435/iteration-through-rows-of-a-dataframe-within-group-of-columns-in-r

标签

dataframe

loops

iteration