Question

我下面的代码为我提供了股票的时间序列结果，并将所有内容分为“买”和“卖”两个类别（基于高于或低于开盘价的收盘价）。

library(dplyr)
library(data.table)
library(quantmod)
library(zoo)

# enter tickers to download time-series data
e <- new.env()
getSymbols("SBUX", env = e)
pframe <- do.call(merge, as.list(e))
#head(pframe)


# get a subset of data
df = pframe$SBUX.Close

colnames(df)[1] <- "Close"
head(df)

# Assign groupings
addGrps <- transform(df,Group = ifelse(Close < lead(Close), "S", "B"))


# create subsets
buys <- addGrps[addGrps$Group == 'B',]
sells <- addGrps[addGrps$Group == 'S',]

现在，我正在尝试将结果按每日利润（Diff）和亏损进行分组，以求出两者的累计总和（利润和亏损）。

我认为应该是这样，但是有些问题了，我不确定是什么。

# find daily differences
df <- df %>%
mutate(Diff = addGrps$Close - lead(addGrps$Close))

# get up and down price movements
ups <- filter(df, Diff > 0 )
downs <- filter(df, Diff <= 0 )

# cumulative sums of longs and shorts
longs<-cumsum(ups$Diff)
shorts<-cumsum(downs$Diff)

Answer 1

我不确定我是否完全关注您的问题/问题，而且似乎有一些不必要的代码。例如，所有这些软件包都是不需要的（至少现在还不需要），目前尚不清楚为什么需要两个用于买卖的子数据框。至少，以下内容将清理您到目前为止所做的一些工作，并在易于使用的数据框架中获取数据。经过澄清，也许这是一个开始。

library(quantmod)
library(tidyverse) # rather than just dplyr

# pull the SBUX data as a data frame and create the necessary new columns:
df <- data.frame(getSymbols(Symbols = 'SBUX', env = NULL)) %>%  # pull the raw data
  rownames_to_column('date') %>%                                # convert the row index to a column
  select(date, close = SBUX.Close) %>%                          # select only the SBUX.Close column and rename it
  mutate(group = ifelse(close < lead(close), 's', 'b')) %>%     # assign the sell or buy group
  mutate(diff = close - lead(close)) %>%                        # create the diff calculation
  mutate(movement = ifelse(diff > 0, 'up', 'down')) %>%         # create the movement classification 
  tbl_df()

# just to view the new data frame: 
df %>% head(5)
# A tibble: 5 x 5
  date       close group    diff movement
 <chr>      <dbl> <chr>   <dbl> <chr>   
1 2007-01-03  17.6 s     -0.0200 down    
2 2007-01-04  17.6 b      0.0750 up      
3 2007-01-05  17.6 b      0.0650 up      
4 2007-01-08  17.5 b      0.0750 up      
5 2007-01-09  17.4 b      0.0550 up 

# calculate the sums of the diff by the movement up or down:
df %>% 
  filter(!is.na(movement)) %>%    # this removes the last date from the data - it cannot have a lead closing price
  group_by(movement) %>%          
  summarize(cum_sum = sum(diff))
# A tibble: 2 x 2
  movement cum_sum
  <chr>      <dbl>
1 down       -489.
2 up          455.

试图找到股票多头和空头的累计每日利润

1 个答案: