Question

假设给我一个数据框，其中有多个列是因素，而感兴趣的列是colA。

例如，假设数据框如下所示：

colA | colB | colC | colD
--------------------------
  1  |  'a' |  1   |  2 
  1  |  'b' |  2   |  3
  4  |  'b' |  2   |  4
  2  |  'a' |  3   |  1
  3  |  'a' |  2   |  6
  3  |  'b' |  1   |  6

我想基于与colA的分组依据来总结每一列，但是以一种方式构造它，以使colB，colC，colD的值分布在行上，而colA的值分布在行上。列。也就是说，当colA值为1时，而colA值为2时，我想要colB值的计数（每个colB值的行），依此类推。 colC和colD相同。产生的数据框将如下所示：

        colA_value1 | colA_value2 | colA_value3 | colA_value4
        -----------------------------------------------------
colB_a |     1      |      1      |      1      |      0
colB_b |     1      |      0      |      1      |      1
colC_1 |     1      |      0      |      1      |      0
colC_2 |     1      |      0      |      1      |      1
colC_3 |     0      |      1      |      0      |      0
colD_1 |     0      |      1      |      0      |      0
colD_2 |     1      |      0      |      0      |      0
colD_3 |     1      |      0      |      0      |      0
colD_4 |     0      |      0      |      0      |      1
colD_6 |     0      |      0      |      2      |      0

偏好使用tidyverse软件包。

Answer 1

这可以用很多提迪尔来完成：

library(tidyverse)

df <- data.frame(colA = c(1L, 1L, 4L, 2L, 3L, 3L), 
                 colB = c("a", "b", "b", "a", "a", "b"), 
                 colC = c(1L, 2L, 2L, 3L, 2L, 1L), 
                 colD = c(2L, 3L, 4L, 1L, 6L, 6L))

df %>% 
    gather(key, value, colA) %>% 
    unite(colA, key, value) %>% 
    gather(key, value, -colA) %>% 
    unite(col, key, value) %>% 
    count(colA, col) %>% 
    spread(colA, n, fill = 0)
#> Warning: attributes are not identical across measure variables;
#> they will be dropped
#> # A tibble: 10 x 5
#>    col    colA_1 colA_2 colA_3 colA_4
#>    <chr>   <dbl>  <dbl>  <dbl>  <dbl>
#>  1 colB_a      1      1      1      0
#>  2 colB_b      1      0      1      1
#>  3 colC_1      1      0      1      0
#>  4 colC_2      1      0      1      1
#>  5 colC_3      0      1      0      0
#>  6 colD_1      0      1      0      0
#>  7 colD_2      1      0      0      0
#>  8 colD_3      1      0      0      0
#>  9 colD_4      0      0      0      1
#> 10 colD_6      0      0      2      0

Answer 2

带有reshape2，melt() + dcast()

library(reshape2)

df <- read.table(header=TRUE, text='colA | colB | colC | colD
              1  |  a |  1   |  2
              1  |  b |  2   |  3
              4  |  b |  2   |  4
              2  |  a |  3   |  1
              3  |  a |  2   |  6
              3  |  b |  1   |  6', sep='|')

df2 <- melt(df, id.vars = 'colA')
df2$value <- trimws(df2$value)
df2$colA <- paste('colA_value', df2$colA, sep='')
df2$variable_value <- paste(df2$variable, df2$value, sep='_')
dcast(df2, variable_value~colA, fun=length)

#   variable_value colA_value1 colA_value2 colA_value3 colA_value4
#1          colB_a           1           1           1           0
#2          colB_b           1           0           1           1
#3          colC_1           1           0           1           0
#4          colC_2           1           0           1           1
#5          colC_3           0           1           0           0
#6          colD_1           0           1           0           0
#7          colD_2           1           0           0           0
#8          colD_3           1           0           0           0
#9          colD_4           0           0           0           1
#10         colD_6           0           0           2           0

在R中旋转数据框

2 个答案: