R为多个条件下的一组值创建一个新列

Question

i am not sure if the title is precise enough to describe my problem. I have a data frame with sets 1:n of two groups A and B and possible status Calm , Action1 and Action2 .

  triggers <- c("Action1", "Action2") 
  df <- data.frame(Set    = c(rep(1, 4), rep(2, 4), rep(1, 4)),                                                                             
                   Group  = c(rep("A", 4), rep("A", 4), rep("B", 4)),                                                                     
                   Status = c(rep("Calm",3), "Action1", rep("Calm",3), 
                                "Action2", rep("Calm", 4))) 
Set Group Status
1    A    Calm
1    A    Calm
1    A    Calm
1    A    Action1
2    A    Calm
2    A    Calm
2    A    Calm
2    A    Action2
1    B    Calm
1    B    Calm
1    B    Calm
1    B    Calm

Based on the vector triggers i want to create a new column where the following condition is met.

If in one set (for each group) an Action1 or Action2 occurs than write into a new column the status Action also for Status Calm .

The result in the new column should be c(rep("Action", 8), rep("Calm", 4)) .

I could not solve the problem with the multiple condition in this data frame. Hope, someone can help me and forgives my non-mathematical description for the problem.

Answer 1

This should work:

df %>% dplyr::group_by(Group, Set) %>% 
  do(mutate(. ,result = ifelse(any(Status %in% triggers), "Action", "Calm")))

Source: local data frame [12 x 4]
Groups: var2, var1 [3]
    Set  Group    Status result
   <dbl> <chr>   <chr>  <chr>
1      1     A    Calm Action
2      1     A    Calm Action
3      1     A    Calm Action
4      1     A Action1 Action
5      2     A    Calm Action
6      2     A    Calm Action
7      2     A    Calm Action
8      2     A Action2 Action
9      1     B    Calm   Calm
10     1     B    Calm   Calm
11     1     B    Calm   Calm
12     1     B    Calm   Calm

Answer 2

Here is a data.table solution as requested

library(data.table)
dt1 <- setDT(df)
dt1[, result := ifelse(sum(Status %in% triggers) == 0, "Calm", "Action"), by = .(Group, Set)]

With the data set provided this is roughly 6.5 times as fast as using dplyr

Answer 3

Considering dft as your inut dataframe, you could use dplyr and try:

dft %>% 
  group_by(Group, Set) %>%
  mutate(nc = if_else( Status %in% triggers, 1, 0)) %>%
  mutate(nc = max(nc)) %>%
  mutate(nc2 = if_else(nc == 1, "Action", as.character(Status))) %>%
  select(nc2)

ps the second and third mutate commands could be merged into one.