Apply value replacement using data dictionary-CodePudding

df1=c("blue,green,green", "blue", "green")
dictionary=data.frame(label=c("blue","green"),value=c(1,2))
want=c("1,2,2","1","2")

I am looking to replace some data using a data dictionary. The data (df1) may have multiple entries in one cell, separated by a comma. I have looked at str_replace_all and I was able to do so bystr_replace_all(colors,c("blue"="1","green"="2")) but I was unable to create c("blue"="1","green"="2") using the dictionary data frame. I have over a hundred items to code so hard coding is not an option. Any guidance on how to make that work or another way around would be greatly appreciated!

CodePudding user response：

Create a named vector from dictionary and use str_replace_all

library(dplyr)
library(stringr)
library(tibble)
dictionary %>% 
   mutate(value = as.character(value)) %>%
   deframe %>%
   str_replace_all(df1, .)
#[1] "1,2,2" "1"     "2"

CodePudding user response：

Here is a base R option:

rename <- with(dictionary, setNames(value, label))

lapply(strsplit(df1, ","), \(x) unname(rename[x])) |>
  lapply(\(x) paste(x, collapse = ",")) |> 
  unlist()

[1] "1,2,2" "1"     "2"

CodePudding user response：

Base R Option 1 using nested vapply() 's:

# dict => named integer vector
dict <- setNames(dictionary$value, dictionary$label)

# Loop through and replace values: df1_replaced => character vector
df1_replaced <- vapply(
   df1, 
   function(x){
      y <- strsplit(
         x, 
         ","
      )
      vapply(
         y, 
         function(z){
            toString(
               dict[
                  match(
                     z, 
                     names(dict)
                     )
                  ]
            )
         },
         character(1)
      )
   },
   character(1),
   USE.NAMES = FALSE
)