如何在R语言中创建基于前日期值的百分比变化数据表？

Absolutely! You’ve got two solid approaches to generate your df2 table—using base R functions (no extra packages needed) or the popular dplyr package for a cleaner, pipe-friendly workflow. Let’s break both down:

Base R Method #

If you prefer sticking to core R tools, you can use diff() to compute differences between consecutive values, then divide by the prior row’s value to get percentage change. We’ll prepend an NA to match your requirement for the first row.

Assuming your first column is the date, and all subsequent columns are numeric values you want to transform:

# Copy the original data frame to avoid modifying the original
df2 <- df

# Loop through each numeric column (skip the date column)
for (col_idx in 2:ncol(df2)) {
  # Calculate percentage change: (current value - previous value)/previous value
  pct_change <- c(NA, diff(df2[[col_idx]]) / df2[[col_idx]][-nrow(df2)])
  # Replace the column with the calculated values
  df2[[col_idx]] <- pct_change
}

# Optional: Round to 4 decimal places to match your example (-.0188)
df2 <- df2 %>% mutate(across(-1, ~round(.x, 4)))

dplyr (Tidyverse) Method #

For a more concise, readable approach, use dplyr’s lag() function to access the prior row’s value, paired with across() to apply the calculation to all numeric columns at once.

First, load the package if you haven’t already:

library(dplyr)

# Generate df2 with percentage changes
df2 <- df %>%
  # Apply the calculation to all columns except the date column (replace "date" with your actual date column name)
  mutate(across(-date, ~ (.x - lag(.x)) / lag(.x))) %>%
  # Optional: Round to 4 decimal places
  mutate(across(-date, ~round(.x, 4)))

Quick Tips #

• If your date column isn’t named date, replace -date with either the column index (e.g., -1 for the first column) or the actual column name.
• The formula (.x - lag(.x)) / lag(.x) simplifies to (.x / lag(.x)) - 1—both will give the exact same percentage change result, use whichever is easier for you to parse.
• The round() step ensures you get the 4-decimal precision shown in your example (-.0188).

内容的提问来源于stack exchange，提问作者Henry Walsh