R语言：高效生成DataFrame累计行和列的方法咨询

Hey there! Let's figure out how to generate those cumulative row sum columns in R without writing repetitive rowSums() calls. It's easier than you think, and we'll make sure we dynamically handle column labels too.

First, let's start with a sample DataFrame to work with—this matches the September/October/November example you mentioned:

# Sample input DataFrame
df <- data.frame(
  September = c(10, 20, 30),
  October = c(5, 15, 25),
  November = c(8, 18, 28)
)

Method 1: Fast Base R with apply() and cumsum() #

This is one of the most efficient approaches, especially for larger datasets. Since cumsum() works column-wise by default, we'll transpose the data to compute row-wise cumulative sums, then transpose back:

# Compute row-wise cumulative sums
cumulative_rows <- t(apply(df, 1, cumsum))

# We don't need the first column (it's just the original September values)
cumulative_rows <- cumulative_rows[, -1]

# Dynamically name the new columns (e.g., "Cumulative_October", "Cumulative_November")
colnames(cumulative_rows) <- paste0("Cumulative_", colnames(df)[-1])

# Bind the cumulative columns to the original DataFrame
df_cumulative <- cbind(df, cumulative_rows)

Running this gives you the original columns plus cumulative sums:

• Cumulative_October = September + October
• Cumulative_November = September + October + November

Method 2: Tidyverse Approach with purrr #

If you prefer the tidyverse syntax, purrr::map_dfc() lets you iterate over column indices cleanly, while dynamically generating column names:

library(dplyr)
library(purrr)

df_cumulative_tidy <- df %>%
  bind_cols(
    map_dfc(2:ncol(df), function(col_index) {
      # Calculate row sum for columns 1 to the current index
      sum_vals <- rowSums(.[, 1:col_index])
      # Create a named column using the current month's label
      col_name <- paste0("Cumulative_", colnames(.)[col_index])
      tibble(!!col_name := sum_vals)
    })
  )

The !! operator here lets us use the dynamically generated col_name as the column name in the tibble.

Method 3: Base R Loop (For Explicit Control) #

If you want a more explicit loop (great for learning or custom tweaks), this works too:

# Initialize a list to store cumulative columns
cumulative_cols <- list()

# Iterate from the 2nd column to the last
for (i in 2:ncol(df)) {
  # Compute row sum for columns 1 to i
  col_sum <- rowSums(df[, 1:i])
  # Create a dynamic column name
  col_name <- paste0("Cumulative_", colnames(df)[i])
  # Add to the list
  cumulative_cols[[col_name]] <- col_sum
}

# Bind to original DataFrame
df_cumulative_loop <- cbind(df, cumulative_cols)

Key Notes on Column Labels #

In all these methods, we use colnames(df)[i] to grab the original column label (like "October" or "November") during iteration. We then prepend a prefix (like "Cumulative_") to make the new column names clear and descriptive—you can adjust this prefix to whatever fits your needs!

内容的提问来源于stack exchange，提问作者Daniel Beltran