Question

我的数据格式如下

structure(list(Flag = c(1, 0, 0, 1, 0, 0, 1, 0), variable = c(3, 
8, 6, 7, 1, 4, 3, 6), sale = c(26, 27, 61, 38, 79, 87, 81, 13
)), .Names = c("Flag", "variable", "sale"), row.names = c(NA, 
-8L), class = "data.frame")

我想按如下方式创建输出

structure(list(Flag = c(1, 0, 0, 1, 0, 0, 1, 0), variable = c(3, 
8, 6, 7, 1, 4, 3, 6), sale = c(26, 27, 61, 38, 79, 87, 81, 13
), begin = c(3, -23, -50, 7, -31, -70, 3, -78), end = c(-23, 
-50, -111, -31, -70, -151, -78, -91)), .Names = c("Flag", "variable", 
"sale", "begin", "end"), row.names = c(NA, -8L), class = "data.frame")

其中ne列的开始和结束基于以下algorathim

if flag=1 then 
    begin=variable;
    end=variable-sale;
----------
else
begin=lag(end) ( i.e the previous value of end variable)
end= lag(end)-sale

我想要的是当flag为1时，“begin”的值等于“variable”值，“end”值是“variable-sale”值。对于其他人而言，begin的值是前一行“end”值，而“end”值是（begin-sales）值任何人都可以帮我解决如何在R中实现这个目标吗？

Answer 1

我认为您提供的示例输出不正确，但我会尝试以下操作：

beginEnd <- by(indf, cumsum(indf$Flag), FUN = function(x) {
  out <- Reduce("-", c(x[, "variable"][1], x[, "sale"]), accumulate = TRUE)
  cbind(begin = head(out, -1), 
        end = tail(out, -1))
})
cbind(indf, do.call(rbind, beginEnd))
#   Flag variable sale begin  end
# 1    1        3   26     3  -23
# 2    0        8   27   -23  -50
# 3    0        6   61   -50 -111
# 4    1        7   38     7  -31
# 5    0        1   79   -31 -110
# 6    0        4   87  -110 -197
# 7    1        3   81     3  -78
# 8    0        6   13   -78  -91

基于先前的行值形成新变量

1 个答案: