加速/替换数百万数据的循环:判断多个日期范围

Question

晚上好，伙计们，我有六百万的数据，他们有四种类型。

z=structure(list(date = structure(c(11866, 16190, 14729, 11718), class = "Date"), 
           beg1 = structure(c(12264, 12264, 13970, 12264), class = "Date"), 
           end1 = structure(c(17621, 14760, 14760, 13298), class = "Date"), 
           ID1 = c(1003587, 1000396, 1010743, 1002113), beg2 = structure(c(NA, 
                                                                           14790, 14790, 13299), class = "Date"), end2 = structure(c(NA, 
                                                                                                                                     17621, 15217, 13969), class = "Date"), ID2 = c(NA, 1024488, 
                                                                                                                                                                                    1027877, 1002824), beg3 = structure(c(NA, NA, 15218, 13970
                                                                                                                                                                                    ), class = "Date"), end3 = structure(c(NA, NA, 17621, 14760
                                                                                                                                                                                    ), class = "Date"), ID3 = c(NA, NA, 1031361, 1002113), beg4 = structure(c(NA, 
                                                                                                                                                                                                                                                              NA, NA, 14790), class = "Date"), end4 = structure(c(NA, NA, 
                                                                                                                                                                                                                                                                                                                  NA, 17621), class = "Date"), ID4 = c(NA, NA, NA, 1021290), 
           realID = c(NA, NA, NA, NA)), row.names = c(267365L, 193587L, 
                                                      5294385L, 2039421L), class = "data.frame")

我试着根据他们的日期来判断和分配一个合适的ID (使用循环)。

for(i in 1:nrow(z)){tryCatch({print(i)
if(between(z$date[i],z$beg1[i],z$end1[i])==T){z$realID[i]=z$ID1[i]}
if(between(z$date[i],z$beg2[i],z$end2[i])==T){z$realID[i]=z$ID2[i]}
if(between(z$date[i],z$beg3[i],z$end3[i])==T){z$realID[i]=z$ID3[i]}
if(between(z$date[i],z$beg4[i],z$end4[i])==T){z$realID[i]=z$ID4[i]}},error=function(e){})}          

密码有效。但是，现在的问题是我有太多的数据，循环效率很低，可能要花将近一天的时间循环。

有人知道如何改进或替换代码吗？非常感谢你。

Answer 1

由于R是一种矢量化语言，为了加快这段代码的速度，最好是对整个向量进行操作，而不是循环遍历每个元素。

简单的解决方案是使用一系列ifelse语句。

z$realID <- ifelse(!is.na(z$beg1) & z$date> z$beg1 & z$date< z$end1, z$ID1, z$realID)
z$realID <- ifelse(!is.na(z$beg2) & z$date> z$beg2 & z$date< z$end2, z$ID2, z$realID)
z$realID <- ifelse(!is.na(z$beg3) & z$date> z$beg3 & z$date< z$end3, z$ID3, z$realID)
z$realID <- ifelse(!is.na(z$beg4) & z$date> z$beg4 & z$date< z$end4, z$ID4, z$realID)

当if语句计算为TRUE时，realID将更新(如果不是)，它将保留其先前的值。