r - 如何在data.table中选择一个值？

Question

我的数据是以下形式

structure(list(atp = c(1, 0, 1, 0, 0, 1), len = c(2, NA, 3, NA, 
NA, 1), Day_1 = c(8, 7, 8, 9, 6, 6), Day_2 = c(94, 94, 102, 97, 
102, 100), Day_3 = c(104, 162, 133, 142, 96, 122)), .Names = c("atp", 
"len", "Day_1", "Day_2", "Day_3"), row.names = c(NA, -6L), class = "data.frame")

我想获得以下输出

structure(list(atp = c(1, 0, 1, 0, 0, 1), len = c(2, NA, 3, NA, 
NA, 1), Day_1 = c(8, 7, 8, 9, 6, 6), Day_2 = c(94, 94, 102, 97, 
102, 100), Day_3 = c(104, 162, 133, 142, 96, 122), output = c(94, 
NA, 133, NA, NA, 6)), .Names = c("atp", "len", "Day_1", "Day_2", 
"Day_3", "output"), row.names = c(NA, -6L), class = "data.frame")

这基本上取决于第 2 列的值，它从第 3、4 或 5 列中选择值。

我是通过以下代码实现的

result<-cbind(y, output=apply(y, 1, function(r) r[r["len"]+2]))

但是这个过程非常耗时。有什么办法可以加快这个过程吗？我如何为此使用 data.tables？

Answer 1

一种可能的方法:

result <- cbind(y, 
                output = unlist(y[3:5])[nrow(y) * (y$len -1) + seq.int(nrow(y))])

另一个(这应该更快):

result <- cbind(y, output = y[3:5][cbind(seq.int(nrow(y)), y$len)])

这两种方法都会导致:

#   atp len Day_1 Day_2 Day_3 output
# 1   1   2     8    94   104     94
# 2   0  NA     7    94   162     NA
# 3   1   3     8   102   133    133
# 4   0  NA     9    97   142     NA
# 5   0  NA     6   102    96     NA
# 6   1   1     6   100   122      6