R data.table 通过引用添加具有其他列值的新列

Question

我有一个示例 data.table 如下：

> dt = data.table("Label" = rep(LETTERS[1:3], 3),
+                 "Col_A" = c(2,3,5,0,2,7,6,8,9),
+                 "Col_B" = c(1,4,3,5,2,0,7,5,8),
+                 "Col_C" = c(2,0,4,1,5,6,7,3,0))
> dt[order(Label)]

 Label Col_A Col_B Col_C
1:     A     2     1     2
2:     A     0     5     1
3:     A     6     7     7
4:     B     3     4     0
5:     B     2     2     5
6:     B     8     5     3
7:     C     5     3     4
8:     C     7     0     6
9:     C     9     8     0

我想创建一个新列，它根据标签列从现有列中获取值。我想要的示例输出如下：

 Label Col_A Col_B Col_C Newcol
1:     A     2     1     2      2
2:     A     0     5     1      0
3:     A     6     7     7      6
4:     B     3     4     0      4
5:     B     2     2     5      2
6:     B     8     5     3      5
7:     C     5     3     4      4
8:     C     7     0     6      6
9:     C     9     8     0      0

逻辑是Newcol值引用基于Label列的各个列。例如Label列的前3行是A，那么Newcol列的前3行就是指Col_A列的前3行。

我试过使用代码 dt[, `:=` ("Newcol" = eval(as.symbol(paste0("Col_", dt$Label))))] 但它没有给出所需的输出。

Answer 1

我们可以使用 kit 包的矢量化开关函数，它像 data.table 一样是 fastverse.

的一部分

dt[, "Newcol" := kit::vswitch(Label, c("A", "B", "C"), list(Col_A, Col_B, Col_C))]

# or if you want to pass column indices
dt[, "Newcol" := kit::vswitch(Label, c("A", "B", "C"), dt[,2:4])]

dt
   Label Col_A Col_B Col_C Newcol
1:     A     2     1     2      2
2:     A     0     5     1      0
3:     A     6     7     7      6
4:     B     3     4     0      4
5:     B     2     2     5      2
6:     B     8     5     3      5
7:     C     5     3     4      4
8:     C     7     0     6      6
9:     C     9     8     0      0

Answer 2

library(data.table)
dt = data.table("Label" = rep(LETTERS[1:3], 3),
                "Col_A" = c(2,3,5,0,2,7,6,8,9),
                "Col_B" = c(1,4,3,5,2,0,7,5,8),
                "Col_C" = c(2,0,4,1,5,6,7,3,0))

dt[, new := ifelse(Label == "A", Col_A, NA)]
dt[, new := ifelse(Label == "B", Col_B, new)]
dt[, new := ifelse(Label == "C", Col_C, new)]

Answer 3

如果您能够使用 dplyr 库，我会使用那里的 case_when 函数。

dt$newCol <- case_when(dt$Col_A == 'A' ~ Col_A, dt$Col_A == 'B' ~ Col_B, dt$Col_A == 'C' ~ Col_C)

我还没有测试过那个代码，但应该是这样的。

Answer 4

与fcase:

cols <- unique(dt$Label)
dt[,newCol:=eval(parse(text=paste('fcase(',paste0("Label=='",cols,"',Col_",cols,collapse=','),')')))][]

    Label Col_A Col_B Col_C newCol
   <char> <num> <num> <num>  <num>
1:      A     2     1     2      2
2:      B     3     4     0      4
3:      C     5     3     4      4
4:      A     0     5     1      0
5:      B     2     2     5      2
6:      C     7     0     6      6
7:      A     6     7     7      6
8:      B     8     5     3      5
9:      C     9     8     0      0

R data.table 通过引用添加具有其他列值的新列

R data.table add new column with values from other columns by referencing

eval

r

data.table