从数据帧中的公共ID求和值

Question

我有一个数据框,看起来像:

df<-data.frame(id=c("xx33","xx33","xx22","xx11","xx11","xx00"),amount=c(10,15,100,20,10,15),date=c("01/02/2013","01/02/2013","02/02/2013","03/03/2013","03/03/2013","04/04/2013"))

    id  amount  date
1   xx33    10  01/02/2013
2   xx33    15  01/02/2013
3   xx22    100 02/02/2013
4   xx11    20  03/03/2013
5   xx11    10  03/03/2013
6   xx00    15  04/04/2013

我想编译所有公共ID并总结数量以及id的出现次数,但是还要携带公共信息,例如每个id(以及任何其他变量)的日期相同.所以,我希望输出为:

    id  sum date        number
1   xx33    25  01/02/2013   2
2   xx22    100 02/02/2013   1
3   xx11    30  03/03/2013   2
4   xx00    15  04/04/2013   1

我试过了

ddply(.data = df, .var = "id", .fun = nrow)

并返回总出现次数,但我无法找到一种方法来汇总所有常见ID而不进行循环.

Answer 1

使用data.table图书馆 -

library(data.table)
dt <- data.table(df)
dt2 <- dt[,list(sumamount = sum(amount), freq = .N), by = c("id","date")]

输出:

> dt2
     id       date sumamount freq
1: xx33 01/02/2013        25    2
2: xx22 02/02/2013       100    1
3: xx11 03/03/2013        30    2
4: xx00 04/04/2013        15    1

Answer 2

以下是使用plyr包的解决方案:

   library(plyr)
ddply(df,.(date,id),summarize,sum=sum(amount),number=length(id))
            date   id sum number
    1 01/02/2013 xx33  25      2
    2 02/02/2013 xx22 100      1
    3 03/03/2013 xx11  30      2
    4 04/04/2013 xx00  15      1

Answer 3

这是R base解决方案

> cbind(aggregate(amount~id+date, sum, data=df), table(df$id))[, -4]
    id       date amount Freq
1 xx33 01/02/2013     25    1
2 xx22 02/02/2013    100    2
3 xx11 03/03/2013     30    1
4 xx00 04/04/2013     15    2