tibbletime获取前一日收盘价及DataFrame每日计算技术咨询

Hey there! No worries about not sharing code/data yet—let's walk through exactly how to handle this daily calculation workflow with tibbletime. Here's a step-by-step breakdown tailored to your needs:

Step 1: Prep Your Environment & Data #

First, make sure you've got the necessary packages loaded. You'll need tibbletime for time-series manipulation and dplyr for data wrangling:

library(tibbletime)
library(dplyr)

Next, ensure your date column is formatted as a date type (critical for tibbletime to work properly). If it's not already, convert it:

# Replace "date_col" with your actual date column name
your_df$date_col <- as.Date(your_df$date_col)

Then convert your regular DataFrame into a tibbletime object—this lets you leverage time-specific functions:

time_df <- as_tbl_time(your_df, index = date_col)

Pro tip: Always sort your data by date first to avoid off-by-one errors with lagged values:

time_df <- time_df %>% arrange(date_col)

Step 2: Get the Previous Day's Close Price #

Tibbletime plays nicely with dplyr's lag() function, which is perfect for grabbing the prior day's CloseUnits. Add this as a new column to your data:

time_df <- time_df %>%
  mutate(prev_day_close = lag(CloseUnits, n = 1))

• The n = 1 argument tells it to pull the value from 1 period back (which is 1 day, since your data is daily).
• Note: The first row will have an NA for prev_day_close (since there's no prior day). You can handle this with replace_na() if needed (e.g., set it to CloseUnits itself for the first day, or 0, depending on your business logic):

time_df <- time_df %>%
  mutate(prev_day_close = replace_na(prev_day_close, CloseUnits))

Step 3: Perform Your Daily Calculations #

Now you can use the prev_day_close column to run whatever daily calculations you need. Let's cover a few common examples based on your fields:

Example 1: Calculate Daily Position Change #

time_df <- time_df %>%
  mutate(daily_position_change = CloseUnits - prev_day_close)

Example 2: Calculate Net Units Traded + Position Validation #

time_df <- time_df %>%
  mutate(net_traded_units = BuyUnits - SellUnits,
         daily_position_change = CloseUnits - prev_day_close,
         # Verify that position change matches net traded units (adjust if other factors apply)
         position_validation = daily_position_change - net_traded_units)

Example 3: Calculate Daily Return Including Interest #

time_df <- time_df %>%
  mutate(daily_return = (CloseUnits - prev_day_close)/prev_day_close + Interest)

Key Notes #

• If your data has gaps (e.g., missing weekends/holidays), tibbletime's pad_time() function can fill in those gaps with NA values, ensuring your lagged values pull the actual prior trading day's close instead of just the calendar day. Use it like this:

time_df <- time_df %>% pad_time(unit = "day")

• Always double-check the sorted order of your dates—if they're out of order, lag() will pull the wrong values!

内容的提问来源于stack exchange，提问作者M du Toit