R语言入门：如何将字符型变量driver重编码为driverrec

Hey there! Let's sort out this recoding problem for you. As an R beginner, it's totally common to get confused by the different syntax rules for recoding functions—let's break down what went wrong and give you a few straightforward solutions.

First, let's clarify the issue: you want to take your driver character variable (with values market, legislation, technology, mixed) and create a new driverrec variable where only market stays as-is, and all other categories are grouped into others.

Why Your Original Code Failed #

It looks like you might be mixing up the syntax for different recode functions (there are a few in R, from packages like dplyr and car). Your attempt passed multiple separate strings as arguments, but most recode functions expect either:

• A single string with all mapping rules separated by semicolons (like car::recode), or
• Key-value pairs directly in the function call (like dplyr::recode).

Let's jump into working solutions:

Solution 1: Use Base R (No Packages Needed) #

This is the simplest approach if you don't want to load extra packages. Use ifelse() to check for the market value and assign others to everything else:

# Create the new variable directly in your data frame
data$driverrec <- ifelse(data$driver == "market", "market", "others")

This works because ifelse() tests each element of data$driver: if it matches "market", it returns "market"; otherwise, it returns "others".

Solution 2: Use dplyr::recode (Tidyverse Style) #

If you're using the tidyverse (a popular set of R packages for data manipulation), dplyr::recode is intuitive. You can list each mapping as a key-value pair:

library(dplyr)

# Option 1: Explicitly map each category
data$driverrec <- recode(data$driver,
                        "market" = "market",
                        "legislation" = "others",
                        "technology" = "others",
                        "mixed" = "others")

# Option 2: More concise (great if you have many categories to group)
# Create a named vector where keys are old values, values are new ones
mapping <- c(legislation = "others", technology = "others", mixed = "others", market = "market")
data$driverrec <- recode(data$driver, !!!mapping)

The !!! here tells dplyr to unpack the named vector into individual key-value pairs.

Solution 3: Use dplyr::case_when (Flexible for Complex Rules) #

If you ever need more complex recoding logic later, case_when is extremely versatile. For your current problem, it looks like this:

library(dplyr)

data$driverrec <- case_when(
  data$driver == "market" ~ "market",  # Keep market as-is
  TRUE ~ "others"                      # All other cases go to others
)

This reads like plain English: "When driver is 'market', use 'market'; for everything else, use 'others'."

Solution 4: Use car::recode (If You Prefer This Package) #

If you want to stick with the recode function you referenced, make sure you're using car::recode and pass all rules in a single string separated by semicolons:

library(car)

data$driverrec <- recode(data$driver, "'market'='market'; 'legislation'='others'; 'technology'='others'; 'mixed'='others'")

Note the single quotes around each value and semicolons separating each rule—this is the required syntax for car::recode.

All these methods will give you exactly the driverrec variable you need. I recommend starting with the base R ifelse() or dplyr::case_when since they're easier to read and debug as a beginner.

内容的提问来源于stack exchange，提问作者dholzer