Change values in a variable based on a conditional value in R-CodePudding

I want to change values in my username variable but only when they meet a condition set from the variable chatforum. For example, I want all instances of users called "Alex" from Canadian chatrooms to be relabeled as "AlexCA":

# mock dataset
library(tidyverse)
username <- c("Alex", "Alex", "Alex", "Alex")
id <- c(1001, 1002, 1003, 1001)
chatforum <- c("Canada", "U.S.", "U.K.", "Canada")

df <- cbind(username, id, chatforum)
df <- as_tibble(df)
glimpse(df)

df <- df  %>% filter(chatforum=="Canada") %>% 
  mutate(username = replace(username, username == "Alex", "AlexCA"))

Though the code above works, I want the entire dataset returned to me, with the changes I just made. Using filter returns a dataset with only the filtered rows, not the entire dataset.

I was advised to use if_else or case_when() but this also changes the username Alice to AlexCA, when I only want the username "Alex" to change when chatroom == Canada:

df <- df %>% mutate(username = if_else(chatforum=="Canada", "AlexCA", username))

Do you know how I can change the values in my username column based on the condition that the value is Alex and the chatroom value is equal to Canada?

CodePudding user response：

For using case_when or ifelse, you can have multiple conditions that must be met in order to the apply the change. So, if chatforum == "Canada" & username == "Alex", then we change the name to AlexCA.

library(tidyverse)

df %>%
  mutate(username = case_when(
    chatforum == "Canada" & username == "Alex" ~ "AlexCA",
    TRUE ~ username
  ))

Or in base R:

df[df$chatforum == "Canada" & df$username == "Alex",]$username <- "AlexCA"

Output

  username id    chatforum
  <chr>    <chr> <chr>    
1 AlexCA   1001  Canada   
2 Alex     1002  U.S.     
3 Alex     1003  U.K.     
4 AlexCA   1001  Canada

But if you need to do this for a lot of countries, then you might want to create a key or add a new column with the abbreviation you want. For example, you could do something like this, where we create an abbreviation from the chatforum, then combine it with the username.

df %>%
  mutate(abrv = toupper(substr(str_replace_all(chatforum, "[[:punct:]]", ""), 1, 2))) %>%
  unite(username, c(username, abrv), sep = "")

#  username id    chatforum
#  <chr>    <chr> <chr>    
#1 AlexCA   1001  Canada   
#2 AlexUS   1002  U.S.     
#3 AlexUK   1003  U.K.     
#4 AlexCA   1001  Canada

Or instead of uniting after creating an abbreviation column, you could still use case_when for certain conditions.

df %>%
  mutate(abrv = toupper(substr(str_replace_all(chatforum, "[[:punct:]]", ""), 1, 2))) %>%
  mutate(username = case_when(
    chatforum == "Canada" & username == "Alex" ~ paste0(username, abrv),
    TRUE ~ username
  ))

#  username id    chatforum abrv 
#  <chr>    <chr> <chr>     <chr>
#1 AlexCA   1001  Canada    CA   
#2 Alex     1002  U.S.      US   
#3 Alex     1003  U.K.      UK   
#4 AlexCA   1001  Canada    CA