如何在R语言中将bcmort数据集的cohort变量拆分为Period和Area两个新变量

Hey there! Let's figure out how to split the cohort variable in the bcmort dataset (from the ISwR package) into two new variables: Period and Area. Since we know exactly the four levels of cohort, we have a few reliable ways to do this in R Studio.

Method 1: Tidyverse (dplyr + tidyr) – Intuitive and Concise #

If you're already using the tidyverse ecosystem, this approach is super straightforward. We'll use string detection to pull out the key terms ("Historical" for period, "National" for area):

# Load required packages
library(ISwR)
library(tidyverse)

# Load the dataset
data(bcmort)

# Split cohort into Period and Area
bcmort <- bcmort %>%
  mutate(
    # Assign Period: "Historical" if the cohort starts with that term, else "Current"
    Period = case_when(
      str_detect(cohort, "^Historical") ~ "Historical",
      TRUE ~ "Current"
    ),
    # Assign Area: "National" if the cohort includes that term, else "Study"
    Area = case_when(
      str_detect(cohort, "National") ~ "National",
      TRUE ~ "Study"
    )
  )

Method 2: Base R – No Extra Packages Needed #

If you prefer sticking to base R, you can split the cohort strings and check for the key words directly:

library(ISwR)
data(bcmort)

# Convert cohort to character and split into individual words
cohort_split <- strsplit(as.character(bcmort$cohort), " ")

# Extract Period by checking for "Historical" in the split parts
bcmort$Period <- sapply(cohort_split, function(parts) {
  if ("Historical" %in% parts) "Historical" else "Current"
})

# Extract Area by checking for "National" in the split parts
bcmort$Area <- sapply(cohort_split, function(parts) {
  if ("National" %in% parts) "National" else "Study"
})

Method 3: Manual Factor Mapping – Precise and Reliable #

Since cohort is a factor with fixed levels, we can create direct mappings for each level. This avoids any string matching edge cases:

library(ISwR)
data(bcmort)

# Create a mapping for Period
period_mapping <- c(
  "Study Group" = "Current",
  "National study group" = "Current",
  "Historical Study Group" = "Historical",
  "Historical National Study Group" = "Historical"
)

# Create a mapping for Area
area_mapping <- c(
  "Study Group" = "Study",
  "National study group" = "National",
  "Historical Study Group" = "Study",
  "Historical National Study Group" = "National"
)

# Apply mappings to create new variables
bcmort$Period <- period_mapping[as.character(bcmort$cohort)]
bcmort$Area <- area_mapping[as.character(bcmort$cohort)]

# Optional: Convert new variables to factors (if you want categorical data)
bcmort$Period <- as.factor(bcmort$Period)
bcmort$Area <- as.factor(bcmort$Area)

Verify the Result #

After running any of these methods, you can double-check the new variables with:

head(bcmort)

This will show you the first few rows of the dataset, including your new Period and Area columns.

内容的提问来源于stack exchange，提问作者Maximus