具有条件的逆变量

Question

我有一个简单的问题，但不知何故无法找出解决方案......

以下是数据集的示例：

dt = data.table(A=rep(c(1:2), each = 5), B = c(1,1,2,2,3,1,2,3,3,1), C =c("a","b","b","b","b","b","a","b","a","a"))

基本上，我想要一个计数器计数器变量，它只在满足条件时重复它的值。条件是A中的后续行应该相同，在B中它们应该不同，在C中它们应该是相同的。这是所需的输出：

dt = data.table(A=rep(c(1:2), each = 5), B = c(1,1,2,2,3,1,2,3,3,1), C =c("a","b","b","b","b","b","a","b","a","a"), counter = c(1,2,2,3,3,4,5,6,7,7))

如您所见，计数器变量仅在满足这些条件时重复它的值。

谢谢！

Answer 1

在逻辑条件下使用cumsum。

n <- nrow(dt)
dt$D <- c(1L, !c(dt$A[-n] == dt$A[-1] & dt$B[-n] != dt$B[-1] & dt$C[-n] == dt$C[-1]))
dt$D <- cumsum(dt$D)

Answer 2

以下是使用shift

中data.table的选项

dt[, counter := cumsum(c(TRUE, !Reduce(`&`, c(Map(`==`, .SD,  
        shift(.SD)),  list(B != shift(B))))[-1])), .SDcols = c(1, 3)]