各组之间的差异

Question

我有一个像这样的数据框：

df = data.frame(
  repetition = c(1,1,1,2,2,2),
  group = c("a", "b", "c", "a", "b", "c"),
  res = c(10,8,9,12,9,10)
)

  repetition group res
1          1     a  10
2          1     b   8
3          1     c   9
4          2     a  12
5          2     b   9
6          2     c  10

现在，对于每次重复，我想计算每组的差异（在

res

中）。所以a组和b组、a组和c组、b组和c组之间的差异。然而，这可能有更多的群体。

我想要这样的输出：

  repetition diff_a_b diff_a_c diff_b_c
1          1        2        1       -1
2          2        3        2       -1

我想要像

diff_a_b

、

diff_a_c

这样的专栏。 dplyr（或其他任何）有办法做到这一点吗？

Answer 1

或许你可以尝试一下

df %>%
    summarise(diff = list(setNames(
        as.data.frame(t(-combn(res, 2, diff))),
        paste0("diff_", combn(group, 2, paste0, collapse = "_"))
    )), .by = repetition) %>%
    unnest(diff)

这给出了

# A tibble: 2 × 4
  repetition diff_a_b diff_a_c diff_b_c
       <dbl>    <dbl>    <dbl>    <dbl>
1          1        2        1       -1
2          2        3        2       -1

Answer 2

我们可以通过

repetition

split，然后在所有

diff

ations 上运行

combn

。

> t(sapply(split(df$res, df$repetition), \(x) -combn(x, 2, FUN=diff)))
  [,1] [,2] [,3]
1    2    1   -1
2    3    2   -1

只是一些化妆品：

> data.frame(repetition=unique(df$repetition),
+            `colnames<-`(t(sapply(split(df$res, df$repetition), \(x) -combn(x, 2, FUN=diff))),
+                         paste0('diff_', combn(unique(df$group), 2, paste, collapse='_'))))
  repetition diff_a_b diff_a_c diff_b_c
1          1        2        1       -1
2          2        3        2       -1

---

数据：

> dput(df)
structure(list(repetition = c(1, 1, 1, 2, 2, 2), group = c("a", 
"b", "c", "a", "b", "c"), res = c(10, 8, 9, 12, 9, 10)), class = "data.frame", row.names = c(NA, 
-6L))