如何在按组忽略 NA 的同时分配索引

Question

给出以下 data.table：

library(data.table)
dt <- data.table(id=c(1,2,3,4), Year1=c(NA, 30, 40, NA), Year2=c(NA, 30, 20 ,70), Year3=c(60, 40, 0, 10), Year4=c(10,20,30,30)) 
dtm <- melt(dt, id.var = 'id')[order(id, variable)]

我想将索引分配给 dtm，这样预期的 return 将如下所示：

   id variable value index
 1:  1    Year1    NA    NA
 2:  1    Year2    NA    NA
 3:  1    Year3    60     1
 4:  1    Year4    10     2
 5:  2    Year1    30     1
 6:  2    Year2    30     2
 7:  2    Year3    40     3
 8:  2    Year4    20     4
 9:  3    Year1    40     1
10:  3    Year2    20     2
11:  3    Year3     0     3
12:  3    Year4    30     4
13:  4    Year1    NA    NA
14:  4    Year2    70     1
15:  4    Year3    10     2
16:  4    Year4    30     3

我试过dtm[ ,index:=1:.N, by = id]，但它不会忽略 NA。

你能给我建议吗？

Answer 1

我们需要指定 'i' 来忽略 NAs

dtm[!is.na(value), index := 1:.N , by = id]
dtm
#    id variable value index
# 1:  1    Year1    NA    NA
# 2:  1    Year2    NA    NA
# 3:  1    Year3    60     1
# 4:  1    Year4    10     2
# 5:  2    Year1    30     1
# 6:  2    Year2    30     2
# 7:  2    Year3    40     3
# 8:  2    Year4    20     4
# 9:  3    Year1    40     1
#10:  3    Year2    20     2
#11:  3    Year3     0     3
#12:  3    Year4    30     4
#13:  4    Year1    NA    NA
#14:  4    Year2    70     1
#15:  4    Year3    10     2
#16:  4    Year4    30     3

或者如果我们使用开发版本（来自@Frank 的评论）

dtm[!is.na(value), z := rowid(id)]

如何在按组忽略 NA 的同时分配索引

How to assign index while ignoring NAs by group

r

na

data.table