Sum of most recent values across groups

Question

For each row of my data I\'d like to compute the sum of most recent value for each group:

dt = data.table(group = c(\'a\',\'b\',\'a\',\         


        

Answer 1

using dplyr, works for many groups, but data must not be data table.

library(dplyr)
library(tidyr)
library(zoo)
dt %>% 
  mutate(row_number = row_number()) %>%
  spread(group, value) %>%
  arrange(row_number) %>%
  mutate_each(funs(na.locf(., na.rm = FALSE))) %>%
  mutate(answer = rowSums(.[,-1:-2], na.rm = T))

Using the above function on example data (notice data.frame() not data.table():

dt = data.frame(group = c('a','b','a','a','b','a'),
                value = c(10, 5, 20, 15, 15, 10),
                desired = c(10, 15, 25, 20, 30, 25))
  desired row_number  a  b answer
1      10          1 10 NA     10
2      15          2 10  5     15
3      25          3 20  5     25
4      20          4 15  5     20
5      30          5 15 15     30
6      25          6 10 15     25

dt = data.frame(group = c('a','b','c','a','a','b','c','a'),
                value = c(10, 5, 20, 25, 15, 15, 30, 10),
                desired = c(10, 15, 35, 50, 40, 50, 60, 55))

  desired row_number  a  b  c answer
1      10          1 10 NA NA     10
2      15          2 10  5 NA     15
3      35          3 10  5 20     35
4      50          4 25  5 20     50
5      40          5 15  5 20     40
6      50          6 15 15 20     50
7      60          7 15 15 30     60
8      55          8 10 15 30     55