循环，子集并将它们放在一起 - R

Question

我有一个数据框。我们就叫他dean_data

dean_data<-data.frame(date=c("23/06/2010", "23/06/2010", "23/06/2010", "29/07/2010", "29/07/2010", "29/07/2010"),
                  hb=c(60, 55, 50, 80, 60, 70),
                  pe=c(11.5, 11.2, 11.7, 8.5, 8, 8.25),
                  v.d=c(2.17, 2.65, 3.66, 2.78, 2.71, 2.68))

首先，我想计算参数“n”次的均值和标准差，按日期（因子）改变线条的位置。

只有在我做数据子集之前，我才能这样做，像这样：

jun13<-subset(dean_data, date=="23/06/2010")

B = 1000
df<-matrix(NA,nrow=B)  
for (b in (1:B)){
  df[b]<-mean(sample(jun13$hb, replace=F)/(sample(jun13$pe, replace=F)*sample(jun13$v.d, replace=F)))
}
df

但是我有几个日期（n = 30）......我想学习以自动方式执行此操作，将数据子集分组并应用参数计算的重复。结果，我希望得到一个按日期显示平均值和 SD 的表格。

Answer 1

所以这是一个尝试：

dat = data_frame(date=c("23/06/2010", "23/06/2010", "23/06/2010", 
                        "29/07/2010", "29/07/2010", "29/07/2010"), 
                 hb=c(60, 55, 50, 80, 60, 70), 
                 pe=c(11.5, 11.2, 11.7, 8.5, 8, 8.25), 
                 v.d=c(2.17, 2.65, 3.66, 2.78, 2.71, 2.68))

dl = split(dat, dat$date) # split the data into groups, based on date

# create a helper function for readability
find_mean = function(hb, pe, v.d) {
    sapply(1:1000, function(n) {
        x = sample(hb, replace=F)
        y = sample(pe, replace=F)
        z = sample(v.d, replace=F)
        mean(x / (y * z))
    })
}

# loop through each subset of the data and find the mean
# output is a list of 1000x1 vectors
m = lapply(dl, function(df) {
  find_mean(df$hb, df$pe, df$v.d)
})

# convert the list to a dataframe 
df = as.data.frame(m)
names(df) = names(dl)

head(df)

  23/06/2010 29/07/2010
1   1.760103   3.116560
2   1.770163   3.117997
3   1.767054   3.108493
4   1.784863   3.131723
5   1.799818   3.107862
6   1.770163   3.128762