使用mutate与case_when转换字符型工龄为数值型报错求助

Let's break down what went wrong with your code and walk through the correct approach to keep your original character variable while adding a numeric version.

Why You Got That Error #

The core issues in your code are two key missteps:

1. Inconsistent return types in case_when: Most branches return numeric values (like 0.99, 2, etc.), but the final TRUE ~ "NA" returns a character string. case_when requires all outputs to match the same data type, hence the error saying it expected a double (numeric) vector instead of a character one.
2. Incorrect mutate syntax: You tried comparing df$length_of_service to the case_when result, which doesn't make sense here—you need to assign a new variable name instead. Also, when using dplyr pipes (%>%), you don't need to reference df$ inside functions like mutate; just use the variable name directly.

Correct Implementation #

You don't need to create a separate empty data frame and merge it later—dplyr's mutate lets you add a new variable directly to your original data frame, preserving the existing length_of_service variable automatically.

Here's the fixed code (assuming your original data frame is named df):

library(dplyr)

# Add a numeric version of length_of_service while keeping the original
df <- df %>%
  mutate(
    los_numeric = case_when(
      length_of_service == "< 1 year" ~ 0.99,
      length_of_service == "2 years" ~ 2,
      length_of_service == "3 years" ~ 3,
      length_of_service == "4 years" ~ 4,
      length_of_service == "5 years" ~ 5,
      length_of_service == "6 years" ~ 6,
      length_of_service == "7 years" ~ 7,
      length_of_service == "8 years" ~ 8,
      length_of_service == "9 years" ~ 9,
      length_of_service == "10+ years" ~ 10,
      # Use numeric NA for unmatched values to keep type consistency
      TRUE ~ NA_real_
    )
  )

This code retains your original length_of_service (character type) and adds a new numeric column los_numeric with your converted values.

Optimization Tips #

Your current case_when has a lot of repetitive lines. You can simplify it using string manipulation to extract numeric values instead of writing a line for every year:

library(dplyr)
library(stringr)

df <- df %>%
  mutate(
    los_numeric = case_when(
      length_of_service == "< 1 year" ~ 0.99,
      length_of_service == "10+ years" ~ 10,
      # Extract digits from the string and convert to numeric for all other cases
      TRUE ~ as.numeric(str_extract(length_of_service, "\\d+"))
    )
  )

This works because all your other values follow the pattern "X years"—str_extract(length_of_service, "\\d+") pulls out the number, and as.numeric converts it to a numeric value. This reduces code repetition and makes it easier to maintain if new categories are added later.

Additional best practices:

• Always use NA_real_ (numeric NA) instead of "NA" (character) when working with numeric vectors to avoid type conflicts.
• After running the code, check for NA values in los_numeric to confirm you didn't miss any categories in your case_when logic.

内容的提问来源于stack exchange，提问作者Data Science Analytics Manager