如何在R语言中为股票日数据添加日收益率列（公式value_t/value_t-1-1）

Got it, let's sort out that daily return formula for you—your current code is almost there, but the key issue is how you're referencing the previous day's value.

The Problem with Your Current Code #

Right now, daily / daily - 1 just divides each value by itself (resulting in 1) then subtracts 1, so you end up with a column full of 0s. That's not the daily return you want!

The Correct Approach #

Daily return is calculated as (Current Day Value / Previous Day Value) - 1. To do this in R, you need to shift your daily vector by one position so each element lines up with the prior day's value. Here are two straightforward ways to implement this:

Method 1: Base R (No Extra Packages)

This uses basic vector indexing to create the shifted comparison:

# Read your CSV (note: read.csv is more tailored for CSV files than read.table)
temp <- read.csv("data.csv")
# Extract the daily value column (adjust the index if your column isn't the second one)
daily <- temp[,2]
# Calculate returns: shift the vector to get previous day's values, add NA for the first row (no prior data)
dailyreturn <- c(NA, daily[-1] / daily[-length(daily)] - 1)
# Add the new column to your data frame (using $ is cleaner than cbind here)
temp$dailyreturn <- dailyreturn
# Write the result to a new CSV (row.names=FALSE avoids extra row number column)
write.csv(temp, "daily_return.csv", row.names=FALSE)

• daily[-1] removes the first element of the vector (all values starting from day 2)
• daily[-length(daily)] removes the last element (all values up to day n-1)
• We prepend NA to match the length of the original data frame—this is standard for the first row, since there's no prior day to compare against. If you prefer to set the first return to 0 instead, just replace NA with 0.

Method 2: Using dplyr (Tidyverse Style, More Readable)

If you use the tidyverse ecosystem, the lag() function makes this even simpler:

# Load the dplyr package (install first with install.packages("dplyr") if needed)
library(dplyr)
# Read data and calculate returns in one pipeline
temp <- read.csv("data.csv") %>%
  # Replace "value" with the actual name of your daily value column
  mutate(dailyreturn = value / lag(value) - 1)
# Export the result
write.csv(temp, "daily_return.csv", row.names=FALSE)

• lag(value) automatically creates a version of your value column shifted back by one row, handling the first row's NA for you. This is more readable, especially if you're working with larger data frames or adding multiple columns.

内容的提问来源于stack exchange，提问作者hacker1337ninja