R：在使用data.table时处理时间不匹配

嘿，我正在尝试从使用ddply过渡到使用data.table，我很快就明白了，但是我仍然需要做一些轻微的调整。以下是我尝试使用玩具数据集所做工作的摘要：

比如说，我有两种产品数周的销售数据

    x <- structure(list(week = c(1, 1, 2, 3, 1, 2, 2, 3, 4), product = c("a", 
    "a", "a", "a", "b", "b", "b", "b", "b"), sold = c(10, 15, 20, 
    25, 30, 35, 40, 45, 50)), .Names = c("week", "product", "sold"
    ), row.names = c(NA, -9L), class = c("data.table", "data.frame"
    ), sorted = c("product", "week"))

        week product sold
     1:    1       a   10
     2:    1       a   15
     3:    2       a   20
     4:    3       a   25
     5:    1       b   30
     6:    2       b   35
     7:    2       b   40
     8:    3       b   45
     9:    4       b   50

xx1
产品周V1
1:125
2:125
3:A20
4:a 3 25
5:B130
6:B275
7:B275
8:B345
9:b 450

问题是我不确定如何删除重复的行，即第2行是多余的。此外，我还想在产品未销售的几周内加入NA，即第4周产品的一行值为NA

---

我相信这是一个简单的问题，我知道如何在ddply中做到这一点，但我无法通过搜索找到我想要的。如果有人能帮我或者把我链接到正确的页面，如果这是重复的，那就太好了

以下是如何通过独特的产品和所有周交叉加入，然后对每组进行汇总

> x[CJ(unique(product), 1:4), sum(sold), by=.EACHI]
   product week V1
1:       a    1 25
2:       a    2 20
3:       a    3 25
4:       a    4 NA
5:       b    1 30
6:       b    2 75
7:       b    3 45
8:       b    4 50

---

如果您使用的是data.table version

，重塑形状可能是另一种选择：

require(reshape2); require(data.table)
(dt2 <- dcast.data.table(dt, product ~ week, fun.aggregate = sum, value.var = "sold", fill = NA, drop = FALSE))
#   product  1  2  3  4
# 1:       a 25 20 25 NA
# 2:       b 30 75 45 50
(dt3 <- melt(dt2, id.vars = "product", variable.name = "week", value.name = "sold"))
#    product week sold
# 1:       a    1   25
# 2:       b    1   30
# 3:       a    2   20
# 4:       b    2   75
# 5:       a    3   25
# 6:       b    3   45
# 7:       a    4   NA
# 8:       b    4   50

require（重塑2）；要求（数据表）
（dt2完美！这正是我所想的，我不知道有交叉连接。
require(reshape2); require(data.table)
(dt2 <- dcast.data.table(dt, product ~ week, fun.aggregate = sum, value.var = "sold", fill = NA, drop = FALSE))
#   product  1  2  3  4
# 1:       a 25 20 25 NA
# 2:       b 30 75 45 50
(dt3 <- melt(dt2, id.vars = "product", variable.name = "week", value.name = "sold"))
#    product week sold
# 1:       a    1   25
# 2:       b    1   30
# 3:       a    2   20
# 4:       b    2   75
# 5:       a    3   25
# 6:       b    3   45
# 7:       a    4   NA
# 8:       b    4   50