如何通过列名提取data.frame列并保留列名？求最优方法

Great question! I totally get the annoyance of either losing column names when using df$columnName (since that returns a plain vector where names(df$columnName) is NULL) or having to type string quotes every time with df["columnName"]. Here are the best, most efficient ways to extract columns while keeping their names intact:

1. Base R: Use [, drop = FALSE] #

This is the simplest base R solution. By default, when you extract a single column with df[, "columnName"], R converts it to a vector (thanks to the drop = TRUE default). Setting drop = FALSE forces R to return a 1-column data.frame, which retains the original column name.

Example:

# Create sample data
df <- data.frame(
  name = c("Alice", "Bob", "Charlie"),
  age = c(25, 30, 35)
)

# Extract "age" column with name preserved
age_col <- df[, "age", drop = FALSE]

# Check the result
names(age_col) # Returns "age"
class(age_col) # Returns "data.frame" (not a vector)

This method also works seamlessly with dynamic column names stored in variables:

target_col <- "age"
age_col <- df[, target_col, drop = FALSE]

2. Tidyverse: Use dplyr::select() #

If you’re working with the tidyverse ecosystem, select() from the dplyr package is incredibly clean and readable. You can pass unquoted column names directly, and it returns a data.frame with all original column names intact.

Example:

library(dplyr)

# Extract "age" column
age_col <- select(df, age)

# Verify
names(age_col) # Returns "age"

For dynamic column names, use all_of() to reference the variable safely:

target_col <- "age"
age_col <- select(df, all_of(target_col))

Bonus: select() makes extracting multiple columns at once trivial while preserving all their names (e.g., select(df, name, age)).

Which Should You Use? #

• Stick with the base R [, drop = FALSE] method if you want to avoid external dependencies.
• Go with dplyr::select() if you’re already using tidyverse tools—it’s more intuitive, especially for complex column selections.

Content of the question comes from Stack Exchange, asked by Marcel Gangwisch