使用唯一 id 平均分割 3 个数据帧

Question

我有三个表：upload、pap1、pap2）每个表有 50 列和 15 万行，我想将三个 dataframes 拆分为匹配的多个数据帧（其中每个子集都有最多 1000 行）使用唯一的主键， 例如，subset_upload1 在subset_pap1 和subset_pap2 中必须具有相同的 ID 等等...

 employee_id<-c(1,2,3)
 employee <- c('John','Peter ','Jolie')
 salary <- c(21000, 23400, 26800)
 startdate <- as.Date(c('2010-11-1','2008-3-25','2007-3-14'))

 upload<- data.frame(employee_id,employee, salary, startdate)

 employee_id<-c(1,2,3)
 line_1<-c('address1','address2','address3')
 line_2<-c('address1','address2','address3')
 postcode<-c('postcode1','postcode2','postcode')

 pap1<-data.frame(employee_id,line_1,line_2,postcode)


 age<-c(57,43,23)
 Height<-c(150,170,190)
 gender<-c('M','M','F')
 enddate<-as.Date(c('2020-11-1','2020-3-25','2020-3-14'))

 pap2<-data.frame(employee_id,age,Height,gender,enddate)

我希望的结果是：

    subupload1<-data.frame(employee_id =1,employee = "John",salary=21000,startdate=as.Date('2010-11-1'))
   subpap1_1<-data.frame(employee_id=1,line_1='address1',line_2='address1',postcode='postcode1')
         subpap2_1<-data.frame(age=57,Height=150,gender='M',enddate=as.Date('202011-1'))

Answer 1

我们可以使用while 循环随机抽取n 的唯一ID，并分别从3 个数据帧中对它们进行子集化以创建新的数据帧。

n <- 1 #Number of unique primary key in one dataframe
remaining_ids <- unique(upload$employee_id)
counter <- 1

while(length(remaining_ids) > n) {
   ids <- sample(remaining_ids, n)
   assign(paste0("subupload_", counter), subset(upload, employee_id %in% ids))
   assign(paste0("subpap1_", counter), subset(pap1, employee_id %in% ids))
   assign(paste0("subpap2_", counter), subset(pap2, employee_id %in% ids))
   counter <- counter + 1
   remaining_ids <- setdiff(remaining_ids, ids)
}

assign(paste0("subupload_", counter),subset(upload, employee_id %in% remaining_ids))
assign(paste0("subpap1_", counter), subset(pap1, employee_id %in% remaining_ids))
assign(paste0("subpap2_", counter), subset(pap2, employee_id %in% remaining_ids))

但是，请尝试使用列表来更好地处理/管理数据，而不是使用大量对象污染您的全局环境。

---

如果我们想将所有这些都写成 csv，我们可以使用 write.csv 而不是 assign，例如：

while(length(remaining_ids) > n) {
  ids <- sample(remaining_ids, n)
  write.csv(subset(upload, employee_id %in% ids), paste0("subupload_", counter, ".csv"))
  write.csv(subset(pap1, employee_id %in% ids), paste0("subpap1_", counter, ".csv"))
  write.csv(subset(pap2, employee_id %in% ids), paste0("subpap2_", counter, ".csv"))
  counter <- counter + 1
  remaining_ids <- setdiff(remaining_ids, ids)
}
write.csv(subset(upload, employee_id %in% remaining_ids), paste0("subupload_", counter, ".csv"))
write.csv(subset(pap1, employee_id %in% remaining_ids), paste0("subpap1_", counter, ".csv"))
write.csv(subset(pap2, employee_id %in% remaining_ids), paste0("subpap2_", counter, ".csv"))

Answer 2

upload<-data.frame(employee_id=c(1,2,3),
                   employee=c('John','Peter','Jolie'),
                   salary=c(21000,23400,26800),
                   startdate=as.Date(c('2010-11-1','2008-3-25','2007-3-14')))

pap1<-data.frame(employee_id=c(1,2,3),
                 line_1=c('address1','address2','address3'),
                 line_2=c('address1','address2','address3'),
                 postcode=c('postcode1','postcode2','postcode'))

pap2<-data.frame(employee_id=c(1,2,3),
                 age=c(57,43,23),
                 Height=c(150,170,190),
                 gender=c('M','M','F'),
                 enddate=as.Date(c('2020-11-1','2020-3-25','2020-3-14')))

subupload1<-data.frame(employee_id=1,employee = "John",salary=21000,startdate=as.Date('2010-11-1'))
subpap1<-data.frame(employee_id=1,line_1='address1',line_2='address1',postcode='postcode1')
subpap2<-data.frame(employee_id=1,age=57,Height=150,gender='M',enddate=as.Date('2020-11-1'))

upload[upload$employee_id%in%1,]
upload[upload$employee_id%in%1:2,]
upload[upload$employee_id%in%1:3,]



upload<-upload[order(upload$employee_id),]
pap1<-pap1[order(pap1$employee_id),]
pap2<-pap2[order(pap2$employee_id),]


upload<-data.frame(employee_id=1:150000,
                   employee=sample(c('John','Peter','Jolie'),150000,replace=TRUE),
                   salary=sample(c(21000,23400,26800),150000,replace=TRUE),
                   startdate=sample(as.Date(c('2010-11-1','2008-3-25','2007-3-14')),150000,replace=TRUE))

split_setting<-c()
for(i in 1:(150000/1000))
  split_setting<-c(split_setting,rep(i,1000))

result<-split(upload,split_setting)

result$`1`
nrow(result$`1`)