R中的数据框：将表转换为预定义结构答案

【问题标题】：Data frame in R: converting table into predifined structure [duplicate]R中的数据框：将表转换为预定义结构
【发布时间】：2023-03-03 18:28:01
【问题描述】：

我在 R 中的数据整理有问题。所以我有一个这样的数据框：

        CardID       Date Amount ItemNumber    ItemCode
1  C0100000111 2001-07-19 449.00          1 I0000000808
2  C0100000111 2001-02-20   9.99          1 I0000000622
3  C0100000111 2001-04-27  49.99          1 I0000000284
4  C0100000111 2001-02-20  69.00          1 I0000000488
5  C0100000111 2001-05-17 299.00          1 I0000000595
6  C0100000111 2001-05-19   5.99          1 I0000000078
7  C0100000199 2001-08-20 229.00          1 I0000000783
8  C0100000199 2001-12-29 229.00          1 I0000000783
9  C0100000199 2001-06-28 139.00          1 I0000000537
10 C0100000343 2001-09-07  99.00          1 I0000000532

我想把它转换成这样的结构，

CardID、FirstPurchaseDate、LastPurchaseDate、NumberOrders、NumberSKUs、TotalAmounts

新表中的每一行 CardID 都是唯一的。我怎样才能做到这一点？

根据上面的表格，我预计会有这样的输出

> Ex
       CardID FirstPurchaseDate LastPurchaseDate NumberOrders NumberSKUs TotalAmounts
1 C0100000111        2001-02-20       2001-07-19            6          6       882.97
2 C0100000199        2001-06-28       2001-12-29            3          2       597.00
3 C0100000343        2001-09-07       2001-09-07            1          1        99.00

【问题讨论】：

更多选项见stackoverflow.com/questions/9847054/…

标签： r dataframe

【解决方案1】：

我们可以使用summarise和dplyr分组后的'CardID'

library(dplyr) 
df1 %>% 
    group_by(CardID) %>% 
    summarise(FirstPurchaseDate = first(Date),
              LastPurchaseDate = last(Date),
              NumberOrders = n(), 
              NumberSKUs= n_distinct(ItemCode),
              TotalAmount = sum(Amount) )

【讨论】：

【解决方案2】：

下面是data.table 版本：

library(data.table)

dt <- data.frame(
  CardID = c("C0100000111", "C0100000111", "C0100000111", "C0100000111", "C0100000111", "C0100000111", "C0100000199", "C0100000199", "C0100000199", "C0100000343"),
  Date = as.Date(c("2001-07-19", "2001-02-20", "2001-04-27", "2001-02-20", "2001-05-17", "2001-05-19", "2001-08-20", "2001-12-29", "2001-06-28", "2001-09-07")),
  Amount = c(449, 9.99, 49.99, 69, 299, 5.99, 229, 229, 139, 99),
  ItemNumber = c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L),
  ItemCode = c("I0000000808", "I0000000622", "I0000000284", "I0000000488", "I0000000595", "I0000000078", "I0000000783", "I0000000783", "I0000000537", "I0000000532")
)

# Convert to data.table
setDT(dt)

dt[, .(
  FirstPurchaseDate = min(Date),
  LastPurchaseDate = max(Date),
  NumberOrders = .N,
  NumberSKUs = length(unique(ItemCode)),
  TotalAmounts = sum(Amount)
), by = CardID]

结果：

        CardID FirstPurchaseDate LastPurchaseDate NumberOrders NumberSKUs TotalAmounts
1: C0100000111        2001-02-20       2001-07-19            6          6       882.97
2: C0100000199        2001-06-28       2001-12-29            3          2       597.00
3: C0100000343        2001-09-07       2001-09-07            1          1        99.00

编辑：Akrun 是第一个，所以去找他的答案吧！留下这个仅供data.table 参考。我应该开始使用dplyr 更多...

【讨论】：