我试图根据其他两列中的变量减去一列中的值。它变得更加复杂,因为并非所有“ID”都在之前和之后的所有站点上。这是数据框的示例:
> time <- c("before", "after", "before", "after", "before", "after", "before")
> site <- c("red", "red","red", "red","green", "green", "blue")
> ID <- c('A','A','B','B','B','B','C')
> cover <- c('10','5','10','5','10','5','3')
> df <- data.frame(time, site, ID, cover)
time site ID cover
1 before red A 10
2 after red A 5
3 before red B 10
4 after red B 5
5 before green B 10
6 after green B 5
7 before blue C 3
您可以看到“C”之前只出现在蓝色中。 结果应该是这样的:
site ID diff
1 red A -5
2 red B -5
3 green B -5
4 blue C <NA>
我尝试了在另一个堆栈溢出帖子中找到的类似问题的代码,但我无法让它工作。
df_new <- df %>% group_by(time, site, ID) %>%
summarise(diff = cover[time == c("after")] - cover[time == c("before")])
知道我哪里出错了吗?
谢谢!
答案 0 :(得分:2)
我们可以通过 diff
进行分组,并带有条件来检查 if/else
library(dplyr)
df %>%
group_by(site, ID) %>%
summarise(diff = if(length(cover) == 1) NA
else diff(cover), .groups = 'drop')
-输出
# A tibble: 4 x 3
site ID diff
<chr> <chr> <dbl>
1 blue C NA
2 green B -5
3 red A -5
4 red B -5
OP 的原始代码不起作用,因为其中一个分组在“时间”中没有“之后”和“之前”。我们可能需要为此指定一个 if/else
条件