Styjun

I'm trying to group various values based on a predefined vector and then update a column.

Sample Data

df <- data.frame(ID = 1:5, Type = c("Windows", "Windows Server", "Cat", "Dog", "Eggs"))

it <- c("Windows", "Windows Server")
animal <- c("Cat", "Dog")
food <- c("Eggs")

What I tried but failed

df$Grouping <- gsub(it, "IT", df$Type)

Error: Pattern > 1

Method that works but long-winded

Using dplyr mutate, I'll be able to achieve what I want but it is very long winded as I have multiple elements in a vector.

df %>% mutate(Grouping = ifelse(Type == "Windows", "IT", 
 ifelse ...))

Intended Output

ID Type Grouping
1 1 Windows IT
2 2 Windows Server IT
3 3 Cat Animal
4 4 Dog Animal
5 5 Eggs Food

Thanks!

One option would be to create a list (or a data.frame) for the mappings and then do a left_join

map <- list(
 it = c("Windows", "Windows Server"),
 animal = c("Cat", "Dog"),
 food = c("Eggs"))

library(dplyr) 
df %>% left_join(stack(map), by = c("Type" = "values"))
# ID Type ind
#1 1 Windows it
#2 2 Windows Server it
#3 3 Cat animal
#4 4 Dog animal
#5 5 Eggs food

Create a list of your predefined vectors and then check which element of the list has the items inside the df$Type

mylist = mget(c("animal", "food", "it"))
names(mylist)[max.col(t(sapply(df$Type, function(x) lapply(mylist, function(y) x %in% y))))]
#[1] "it" "it" "animal" "animal" "food"

the question as posted doesn't make a lot of sense. Specifically, with the sample data, Its no simpler to to store the independent type vectors than to store the the type as an attribute of the initial data frame. maybe you could add some color that gives more detail about the nature of the problem.

with that said, assuming your problem is that the lookup vectors are stored in a different source and need to be loaded independently, a simple loop should suffice. (I'm using data.table, cause i don't even remember how to use a raw data.frame anymore):

df <- data.table(ID = 1:5, Type = c("Windows", "Windows Server", "Cat", "Dog", "Eggs"))
it <- c("Windows", "Windows Server")
animal <- c("Cat", "Dog")
food <- c("Eggs")

lookup.names <- c("it", "animal", "food")
for (z in 1:length(lookup.names) ) 
 lookup <- get(lookup.names[z]) #maybe need to do some more sophisticated load, like from a file or database
 df[Type %in% lookup, Grouping := lookup.names[z]]