R 数据帧到4D阵列

我需要将数据帧（

“d”

在下面的虚拟代码中）重塑为4D数组。在特定情况下，我希望下面的变量

“count”

包含在一个4D数组中，该数组的维度为site by year by date by method

d <- data.frame(site=c(rep("aaa", 4), rep("bbb", 4)), date=c(114:117, 220:223), 
                year=c(2005:2008, 2009:2012), count=rpois(8, 34),
                method=c(rep(1, 2), rep(2, 2), rep(1, 1), rep(2, 3)))

d我们可以使用
xtabs（）

资料
d
xtabs(count ~ ., d)
# , , year = 2005, method = 1
# 
# date
# site  114 115 116 117 220 221 222 223
# aaa  45   0   0   0   0   0   0   0
# bbb   0   0   0   0   0   0   0   0
# 
# , , year = 2006, method = 1
# 
# date
# site  114 115 116 117 220 221 222 223
# aaa   0  38   0   0   0   0   0   0
# bbb   0   0   0   0   0   0   0   0
# 
# [...]
# , , year = 2005, method = 2
# 
# date
# site  114 115 116 117 220 221 222 223
# aaa   0   0   0   0   0   0   0   0
# bbb   0   0   0   0   0   0   0   0
# 
# , , year = 2006, method = 2
# 
# date
# site  114 115 116 117 220 221 222 223
# aaa   0   0   0   0   0   0   0   0
# bbb   0   0   0   0   0   0   0   0
# 
# [...]

d <- structure(list(site = structure(c(1L, 1L, 1L, 1L, 2L, 2L, 2L, 
2L), .Label = c("aaa", "bbb"), class = "factor"), date = c(114L, 
115L, 116L, 117L, 220L, 221L, 222L, 223L), year = 2005:2012, 
    count = c(45L, 38L, 35L, 36L, 30L, 42L, 25L, 31L), method = c(1, 
    1, 2, 2, 1, 2, 2, 2)), class = "data.frame", row.names = c(NA, 
-8L))