Question

我想从长到宽投射数据，但将变量显示为行而不是列

数据集：

df1 <- data.frame(Area = c('a', 'a', 'b', 'b', 'c', 'c'),
                  Period =c(1,2,1,2,1,2), 
                  var1 = c(1,2,3,4,5,6),
                  var2 = c(7,8,9,10,11,12),
                  var3 = c(13,14,15,16,17,18)
                  )

Area Period var1 var2 var3
1   a   1   1   7   13
2   a   2   2   8   14
3   b   1   3   9   15
4   b   2   4   10  16
5   c   1   5   11  17
6   c   2   6   12  18

我想去：

             Period 
Area          1   2
 a    Var1      
 a    Var2      
 a    Var3      
 b    Var1      
 b    Var2      
 b    Var3

我尝试过dcast并设法获得：

df1 = dcast(setDT(df1), Area ~ Period ,sum, value.var=c("var1", "var2", "var3"))

Area    var1_1    var1_2    var2_1    var2_2    var3_1    var3_2
1         a         1     2     7         8     13  14
2         b         3     4     9         10    15  16
3         c         5     6     11        12    17  18

..但这不是我真正追求的。任何帮助将不胜感激！

Answer 1

首先融合为长格式，然后重新铸造为宽

library( data.table )

dcast( 
  melt( setDT(df1), 
        id.vars = c("Area", "Period"), 
        measure.vars = patterns( var = "^var") ), 
  Area + variable ~ Period, 
  value.var = "value" )

#    Area variable  1  2
# 1:    a     var1  1  2
# 2:    a     var2  7  8
# 3:    a     var3 13 14
# 4:    b     var1  3  4
# 5:    b     var2  9 10
# 6:    b     var3 15 16
# 7:    c     var1  5  6
# 8:    c     var2 11 12
# 9:    c     var3 17 18

Answer 2

使用dplyr和tidyr，您可以执行以下操作：

df1 %>% 
 pivot_longer(-c(1:2)) %>%
 pivot_wider(names_from = Period, values_from = value)

  Area  name    `1`   `2`
  <fct> <chr> <dbl> <dbl>
1 a     var1      1     2
2 a     var2      7     8
3 a     var3     13    14
4 b     var1      3     4
5 b     var2      9    10
6 b     var3     15    16
7 c     var1      5     6
8 c     var2     11    12
9 c     var3     17    18

使用多个变量在R中投射数据

2 个答案: