r - 使用频率列将宽转换为长

Question

我正在尝试将我的 data.frame 从宽表转换为带有频率列的长表。

data("UCBAdmissions")
ucb_admit <- as.data.frame(UCBAdmissions)
ucb_admit
   Admit    Gender Dept Freq
1  Admitted   Male    A  512
2  Rejected   Male    A  313
3  Admitted Female    A   89
4  Rejected Female    A   19
...  

而且我想收集这些数据(tidyr 包，类似于从 reshape 中熔化)，但使用 Freq 来指定该行应该重复多少次。

因此，我的目标数据看起来像:

     Admit      Gender Dept
1    Admitted   Male    A
2    Admitted   Male    A
3    Admitted   Male    A
4    Admitted   Male    A
5    Admitted   Male    A
6    Admitted   Male    A
...
4523 Rejected Female    F
4524 Rejected Female    F
4525 Rejected Female    F
4526 Rejected Female    F

我想使用 tidyr::gather() 来执行此操作，但是我的结果不正确，因为我不确定是否/如何包含 Freq 列？

谢谢

Answer 1

这看起来不像 gather 的工作因为数据是聚合的，而不是广泛的。您可以通过重复行索引 Freq 来使用索引“分解”数据。每行的次数。以下是使用基础 R 和 dplyr 的方法.

library(dplyr)

# Base R
ucb_admit_disagg = ucb_admit[rep(1:nrow(ucb_admit), ucb_admit$Freq), 
                             -grep("Freq", names(ucb_admit))]

# dplyr
ucb_admit_disagg = ucb_admit %>% 
  slice(rep(1:n(), Freq)) %>% 
  select(-Freq)

这是数据框的一部分。我在输出中添加了省略号以标记行序列中的中断。

ucb_admit_disagg[c(1:6, 510:514, 4523:4526), ]

          Admit Gender Dept
1      Admitted   Male    A
1.1    Admitted   Male    A
1.2    Admitted   Male    A
1.3    Admitted   Male    A
1.4    Admitted   Male    A
1.5    Admitted   Male    A
...
1.509  Admitted   Male    A
1.510  Admitted   Male    A
1.511  Admitted   Male    A
2      Rejected   Male    A
2.1    Rejected   Male    A
...
24.313 Rejected Female    F
24.314 Rejected Female    F
24.315 Rejected Female    F
24.316 Rejected Female    F