循环遍历 r 中的 groupby 列并应用 function

Question

Hello everyone I would need help in order to loop over a dataframe by groups of columns.

Here is an example of dataframe

  Group       Species Values
1    G1 Cattus_cattus   Val1
2    G1 Cattus_cattus   Val2
3    G1 Cattus_cattus   Val3
4    G2   Canis_lupus   Val4
5    G2   Canis_lupus   Val5
6    G3  Griseus_lupa   Val6
7    G4  Griseus_lupa   Val7

I would like to:

1 - loop over c(df$Group,df$Species)

2 - take the df$Values and store it as a vector

3 - put that vector into a function called afunction

4 - open a treefile with anotherfunction where its name is the df$Group name

5 - get the output value of that function and add it into a new_column

So here is an exemple of what the code should do:

first groups is G1,Cattus_cattus :

  Group       Species Values
1    G1 Cattus_cattus   Val1
2    G1 Cattus_cattus   Val2
3    G1 Cattus_cattus   Val3

Then I open the treefile with treefile <- anotherfunction(G1)

Then I generate the output value such as output_value<-afunction(treefile,c("Val1","Val2","Val3))

then the output_value = 30

so I add 30 into the df:

  Group       Species Values new_column
1    G1 Cattus_cattus   Val1 30
2    G1 Cattus_cattus   Val2 30
3    G1 Cattus_cattus   Val3 30

if there is only one row within the Group, then I do nothing and add a NA.

Note that of course it is a nonexisting function, so you cannot reproduce the exemple.

Ath the and we should get something like (where new_column values are random here).

  Group       Species Values new_column
1    G1 Cattus_cattus   Val1 30
2    G1 Cattus_cattus   Val2 30
3    G1 Cattus_cattus   Val3 30
4    G2   Canis_lupus   Val4 21
5    G2   Canis_lupus   Val5 21
6    G3  Griseus_lupa   Val6 NA
7    G4  Griseus_lupa   Val7 NA

Does someone have an idea please? So fare I known how to loop over a dataframe using a for loop but here I do not known how to deal with groups composed of 2 colums..

data

structure(list(Group = structure(c(1L, 1L, 1L, 2L, 2L, 3L, 4L
), .Label = c("G1", "G2", "G3", "G4"), class = "factor"), Species = structure(c(2L, 
2L, 2L, 1L, 1L, 3L, 3L), .Label = c("Canis_lupus", "Cattus_cattus", 
"Griseus_lupa"), class = "factor"), Values = structure(1:7, .Label = c("Val1", 
"Val2", "Val3", "Val4", "Val5", "Val6", "Val7"), class = "factor")), class = "data.frame", row.names = c(NA, 
-7L))

Answer 1

You can try something like this:

library(dplyr)
library(purrr)

df %>%
  group_by(Group) %>%
  summarise(treefile = anotherfunction(first(Group)), 
            Values = list(Values)) %>%
  mutate(new_column = map2_dbl(treefile, Values, afunction)) 

This would give you a summarised dataframe. To get the same number of rows back you can left_join with df by Group .

Answer 2

Here is how you do it:

anotherfunction = function(x){
  #do something with your treefile
  ifelse("Val2" %in% x, 30, ifelse("Val4" %in% x, 21, NA))
}

df %>% 
  group_by(Group) %>% 
  mutate(new_column=anotherfunction(Values))

You did not give a lot of information about anotherfunction() so I used an ugly nested ifelse() to mimic the behavior.

The key is that mutate() will use the Values inside the Groups.

To explore this, you can try to run the code:

anotherfunction = function(x){browser()}