1. R语言运行效率分析(1)
2020-01-31 本文已影响0人
灭绝老衲
测试程序运行所需时间的函数的选择
在R语言中,统计一个程序体运行时间一般采用的函数为Sys.time()
或者为proc.time()
。不过,这两个函数只能根据时间差判断程序执行一次所用的时间,若要重复多次进行平均时间的统计,则显得无能为力。
在此,我们采用microbenchmark
函数包来进行统计程序运行时间。该函数使用很简单,只需要输入待测试代码,并且指定“times=N”,程序就会重复运行代码N次,然后返回运行时间的平均值。默认的话times=100。
注:本研究只从表象上展示运算结果,不进行比如时间复杂度和空间复杂度方面的探讨。想要了解这一部分内容,可以参考以下两篇文章:
生成模拟数据以待测试
#generating n integer data between 1 to 12
month_digital<-function(n){
month_digital<-c()
for (i in 1:n){
month_digital[i]<-sample(1:12,1,replace = F)
}
return(month_digital)
}
运算过程模拟
任务:根据生成的测试数据(1~12)生成对应的月份的英文名字,并判断该月份所属的季节
方法1: 采用 for + if 语句实现
1: 自定义函数
# digital was translated into month's englishname
Month_name_for_if<-function(month){
Month_name<-c()
for (i in 1:length(month)){
if (month[i]==1) Month_name[i]<-"Jan"
if (month[i]==2) Month_name[i]<-"Feb"
if (month[i]==3) Month_name[i]<-"Mar"
if (month[i]==4) Month_name[i]<-"Apr"
if (month[i]==5) Month_name[i]<-"May"
if (month[i]==6) Month_name[i]<-"Jun"
if (month[i]==7) Month_name[i]<-"Jul"
if (month[i]==8) Month_name[i]<-"Aug"
if (month[i]==9) Month_name[i]<-"sep"
if (month[i]==10) Month_name[i]<-"Oct"
if (month[i]==11) Month_name[i]<-"Nov"
if (month[i]==12) Month_name[i]<-"Dec"
}
return(Month_name)
}
# digital was translated into season's englishname
Season_name_for_if<-function(month){
Season_name<-c()
for (i in 1:length(month)){
if (month[i]==1) Season_name[i]<-"Winter"
if (month[i]==2) Season_name[i]<-"Winter"
if (month[i]==3) Season_name[i]<-"Spring"
if (month[i]==4) Season_name[i]<-"Spring"
if (month[i]==5) Season_name[i]<-"Spring"
if (month[i]==6) Season_name[i]<-"Summer"
if (month[i]==7) Season_name[i]<-"Summer"
if (month[i]==8) Season_name[i]<-"Summer"
if (month[i]==9) Season_name[i]<-"Autumn"
if (month[i]==10) Season_name[i]<-"Autumn"
if (month[i]==11) Season_name[i]<-"Autumn"
if (month[i]==12) Season_name[i]<-"Winter"
}
return(Season_name)
}
#generating month and season english
result_for_if<-function(n){
month<-month_digital(n) # n months
Month_name_for_if<-Month_name_for_if(month)# months' names
Season_name_for_if<-Season_name_for_if(month) #seasons' names
df<-data.frame(month,Month_name_for_if,Season_name_for_if)
return(df)
}
2: 调用函数进行运算
month<-month_digital(10) #随机生成10个数据进行测试
microbenchmark::microbenchmark(Month_name_for_if(month))
microbenchmark::microbenchmark(Season_name_for_if(month))
microbenchmark::microbenchmark(result_for_if(month))
Unit: microseconds
expr min lq mean median uq max neval
Month_name_for_if(month) 16.299 16.6255 325.218 16.831 17.174 30813.88 100
Unit: microseconds
expr min lq mean median uq max neval
Season_name_for_if(month) 15.818 16.347 322.3124 16.7195 18.312 30467.87 100
Unit: microseconds
expr min lq mean median uq max neval
result_for_if(month) 846.104 854.45 960.1528 867.8155 881.9025 5631.1 100
There were 50 or more warnings (use warnings() to see the first 50)
(未完!待续……)