r - 如何加速或矢量化 for 循环？

Question

我想通过矢量化或使用 Data.table 或其他方法来提高 for 循环的速度。我必须在 1,000,000 行上运行代码，我的代码真的很慢。

该代码是不言自明的。以防万一，我在下面提供了解释。我已经包括了函数的输入和输出。希望您能帮助我使该功能更快。

我的 目标 是对向量“Volume”进行分箱，其中每个分箱等于 100 份。向量“Volume”包含交易的股票数量。这是它的样子:

head(Volume, n = 60)
[1]  5  3  1  5  3  1  1  1  1  1  1  1 18  1  1 18  2  7 13  2  7 13  3  2  1  1  3  2  1  1  1
[32]  1  6  6  1  1  1  1  1  1  1  1 18  2  1  1  2  1 14 18  2  1  1  2  1 14  1  1  9  5

向量“binIdexVector”与“Volume”长度相同，包含bin号；即前 100 股的每个元素获得数字 1，接下来 100 股的每个元素获得数字 2，接下来 100 股的每个元素获得数字 3，以此类推。这是该向量的样子:

 head(binIdexVector, n = 60)
[1] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
[48] 2 2 3 3 3 3 3 3 3 3 3 3 3

这是我的函数 :

#input as a vector
Volume<-c(5L, 3L, 1L, 5L, 3L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 18L, 1L, 1L, 
                   18L, 2L, 7L, 13L, 2L, 7L, 13L, 3L, 2L, 1L, 1L, 3L, 2L, 1L, 1L, 
                   1L, 1L, 6L, 6L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 18L, 2L, 1L, 
                   1L, 2L, 1L, 14L, 18L, 2L, 1L, 1L, 2L, 1L, 14L, 1L, 1L, 9L, 5L, 
                   2L, 1L, 1L, 1L, 1L, 9L, 5L, 2L, 1L, 1L, 1L, 2L, 1L, 1L, 3L, 1L, 
                   1L, 2L, 1L, 2L, 1L, 1L, 3L, 1L, 1L, 2L, 9L, 9L, 3L, 3L, 1L, 1L, 
                   1L, 1L, 5L, 5L, 8L, 8L, 2L, 1L, 2L, 1L, 10L, 10L, 10L, 10L, 10L, 
                   10L, 10L, 10L, 9L, 9L, 1L, 1L, 8L, 1L, 8L, 1L, 8L, 8L, 2L, 1L, 
                   1L, 1L, 1L, 1L, 1L, 2L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 
                   1L, 1L, 3L, 3L, 3L, 3L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 5L, 5L, 
                   1L, 2L, 7L, 1L, 2L, 7L, 1L, 1L, 1L, 1L, 2L, 1L, 10L, 1L, 1L, 
                   1L, 1L, 1L, 1L, 2L, 1L, 10L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
                   1L, 1L, 30L, 1L, 2L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 3L, 
                   1L, 2L, 2L, 1L, 1L, 1L, 1L, 2L, 2L, 1L, 2L, 2L, 2L, 1L, 2L, 1L, 
                   10L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 1L, 10L, 1L, 1L, 1L, 1L, 1L, 
                   1L, 1L, 1L, 1L, 1L, 30L, 1L, 2L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 
                   1L, 1L, 3L, 1L, 2L, 2L, 1L, 1L, 1L, 1L, 2L, 2L, 1L, 2L, 2L, 2L, 
                   1L, 1L, 1L, 1L, 1L, 1L, 1L, 7L, 7L, 3L, 1L, 1L, 1L, 4L, 3L, 1L, 
                   1L, 1L, 4L, 25L, 1L, 1L, 25L, 1L, 1L, 2L, 1L, 1L, 2L, 1L, 1L, 
                   1L, 1L, 1L, 1L, 1L, 1L, 2L, 1L)

binIdexVector <- numeric(length(Volume))

# initialize 
binIdex <-1
totalVolume <-0

for(i in seq_len(length(Volume))){

  totalVolume <- totalVolume + Volume[i]  

  if (totalVolume <= 100) {

    binIdexVector[i] <- binIdex

  } else {

    binIdex <- binIdex + 1
    binIdexVector[i] <- binIdex
    totalVolume <- Volume[i]
  }
}

# output:
> dput(binIdexVector)
c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 
  1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 
  2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 
  3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 
  3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 
  4, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 6, 6, 6, 6, 
  6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 
  6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 7, 
  7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 
  7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 
  7, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 
  8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 
  8, 8, 8, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 
  9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 10, 10, 10, 10, 10, 10, 10, 
  10, 10, 10, 10, 10, 10, 10, 10, 10, 10)

非常感谢您的帮助!

> sessionInfo()
R version 3.1.2 (2014-10-31)
Platform: x86_64-w64-mingw32/x64 (64-bit)

locale:
[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252   
[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C                          
[5] LC_TIME=English_United States.1252    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

loaded via a namespace (and not attached):
[1] tools_3.1.2

Answer 1

当矢量化困难时，您可以使用 Rcpp。

library(Rcpp)
cppFunction('
  IntegerVector bin(NumericVector Volume, int n) {
    IntegerVector binIdexVector(Volume.size());
    int binIdex = 1;
    double totalVolume =0;

    for(int i=0; i<Volume.size(); i++){
      totalVolume = totalVolume + Volume[i];
      if (totalVolume <= n) {
        binIdexVector[i] = binIdex;
      } else {
        binIdex++;
        binIdexVector[i] = binIdex;
        totalVolume = Volume[i];
      }
    }
    return binIdexVector;
  }')

all.equal(bin(Volume, 100), binIdexVector)
#[1] TRUE

它比 findInterval(cumsum(Volume), seq(0, sum(Volume), by=100)) 快(这当然给出了一个不准确的答案)