如何用quantmod包getSymbols实现年度数据子集化并存储为新对象

Hey there! Let me break down how to use the getSymbols() function from the quantmod package to fetch financial data, subset it to a specific full year (like the previous calendar year), and store it as a new object. I'll cover both manual year specification and automatic previous-year handling, plus fix a small gotcha from your example code.

1. Set Up Required Packages #

First, make sure you have quantmod (for data fetching) and dplyr (for filtering) installed and loaded. If you want to simplify automatic date calculations, lubridate is a handy addition too:

# Install packages if you haven't already
install.packages(c("quantmod", "dplyr", "lubridate"))

# Load the packages
library(quantmod)
library(dplyr)
library(lubridate)

2. Fetch Data with getSymbols() #

The getSymbols() function pulls financial data from sources like Yahoo Finance. Let's walk through your example for Indonesian stock MTDL.JK (weekly data), with notes on key parameters:

• auto.assign = FALSE: Prevents the function from automatically creating a global variable with the ticker name (we want to assign it to mtdl manually)
• src = "yahoo": Specifies Yahoo Finance as the data source
• periodicity = "weekly": Fetches weekly instead of daily data
• na.omit(): Removes any rows with missing values

# Fetch weekly data for MTDL.JK and clean missing values
mtdl <- na.omit(getSymbols("MTDL.JK", auto.assign = FALSE, src = "yahoo", periodicity = "weekly"))

Quick Note: xts Object Structure #

getSymbols() returns an xts object (time-series data), where row names are dates. To use dplyr's filter() (like your example), we first need to convert the row names into a dedicated DATE column:

# Convert xts object to a data frame with a DATE column
mtdl_df <- as.data.frame(mtdl) %>%
  mutate(DATE = index(mtdl))

3. Subset to a Specific Year #

Option 1: Manual Year Specification (e.g., 2018) #

If you know exactly which year you want, use hard-coded start/end dates with filter():

# Subset data to full 2018 calendar year
week.year.mtdl <- mtdl_df %>%
  filter(DATE >= as.Date("2018-01-01") & DATE <= as.Date("2018-12-31"))

Option 2: Automatic Previous Year (No Hard-Coded Dates) #

To dynamically fetch the complete previous calendar year (no need to update dates annually), calculate the start/end dates programmatically:

# Calculate start and end of the previous year
last_year_start <- floor_date(Sys.Date(), "year") - years(1)
last_year_end <- floor_date(Sys.Date(), "year") - days(1)

# Subset to the previous year's data
week.year.mtdl_last_year <- mtdl_df %>%
  filter(DATE >= last_year_start & DATE <= last_year_end)

Bonus: xts Native Subsetting (Simpler!) #

If you prefer working directly with xts objects (no need to convert to a data frame), you can subset by year in one line:

# Directly subset xts object to 2018 data
week.year.mtdl_xts <- mtdl["2018"]

# For previous year, use:
week.year.mtdl_last_year_xts <- mtdl[paste0(year(Sys.Date()) - 1)]

内容的提问来源于stack exchange，提问作者user9092800