为啥 rev(factor) 不能作为反转 dplyr::top_n() 的 wt 参数的一种方式？

Posted 2023-03-31

【中文标题】为啥 rev(factor) 不能作为反转 dplyr::top_n() 的 wt 参数的一种方式？

【英文标题】：Why doesn't rev(factor) work as a way to reverse the wt argument of dplyr::top_n()?为什么 rev(factor) 不能作为反转 dplyr::top_n() 的 wt 参数的一种方式？

【发布时间】：2020-12-17 00:30:43

【问题描述】：

目标：返回 thing1=F 和 thing2=MeFirst 的行

为什么这不起作用？

tibble(
    row = 1:10,
    thing1 = c(rep("F",5),rep("L",5)),
    thing2 = c(rep("MeSecond",4),rep("MeFirst",2),rep("MeSecond",4))
) %>%
    mutate(
        thing1 = factor(thing1, levels = c("F", "L")),
        thing2 = factor(thing2, levels = c("MeFirst", "MeSecond"))
    ) %>%
    top_n(
        .,
        n = 1,
        wt = rev(thing1)
    ) %>%
    top_n(
        .,
        n = 1,
        wt = rev(thing2)
    )

以上返回行 2:5。

我知道这确实有效：

tibble(
    row = 1:10,
    thing1 = c(rep("F",5),rep("L",5)),
    thing2 = c(rep("MeSecond",4),rep("MeFirst",2),rep("MeSecond",4))
) %>%
    mutate(
        thing1 = factor(thing1, levels = c("F", "L")),
        thing2 = factor(thing2, levels = c("MeFirst", "MeSecond"))
    ) %>%
    top_n(
        .,
        n = -1,
        wt = thing1
    ) %>%
    top_n(
        .,
        n = -1,
        wt = thing2
    )

但问题是，为什么 rev(thing2) 不起作用？

【参考方案1】：

简短的回答是您想要反转因子的水平，而不是因子本身：

library(dplyr)
library(forcats)

tibble(row = 1:10,
       thing1 = c(rep("F", 5), rep("L", 5)),
       thing2 = c(rep("MeSecond", 4), rep("MeFirst", 2), rep("MeSecond", 4))) %>%
  mutate(thing1 = factor(thing1, levels = c("F", "L")),
         thing2 = factor(thing2, levels = c("MeFirst", "MeSecond"))) %>%
  top_n(n = 1, wt = fct_rev(thing1)) %>%
  top_n(n = 1, wt = fct_rev(thing2))

# A tibble: 1 x 3
    row thing1 thing2 
  <int> <fct>  <fct>  
1     5 F      MeFirst