R:更难的 pivot_longer 参数

Question

我有一个具有这些名称的数据框:

df <- tribble( ~ state, ~ county, ~ kfr_asian_pooled_p25, ~ kfr_asian_pooled_p75, 
               ~ kfr_black_pooled_p25, ~ kfr_black_pooled_p75, 
               ~ kfr_pooled_pooled_p25, ~ kfr_pooled_pooled_p75, ~ kfr_white_pooled_p25, 
               ~ kfr_white_pooled_p75, ~ kids_count, ~ kfr_asian_pooled_p25_se, 
               ~ kfr_asian_pooled_p75_se, ~ kfr_black_pooled_p25_se, 
               ~ kfr_black_pooled_p75_se, ~ kfr_white_pooled_p25_se, 
               ~ kfr_white_pooled_p75_se,
               1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)

当然，数据集是巨大的，但我想要实现的是利用其最大潜力实现更长时间的转向。我的意思是同时抓取多个变量并旋转它们及其名称。

df <- tribble(~ state, ~ county, ~ kids_count,   ~ race, ~ percentile,   ~ se, ~ value,
                     1,       2,            3,  "asian",        "p25",   TRUE,   value,
                     2,       3,            4,  "black",        "p25",   TRUE,   value,
                     3,       4,            5,  "white",        "p25",   TRUE,   value,
                     1,       2,            3,  "asian",        "p75",   TRUE,   value,
                     2,       3,            4,  "black",        "p75",   TRUE,   value,
                     3,       4,            5,  "white",        "p75",   TRUE,   value,
                     1,       2,            3,  "asian",        "p25",  FALSE,   value,
                     2,       3,            4,  "black",        "p25",  FALSE,   value,
                     3,       4,            5,  "white",        "p25",  FALSE,   value,
                     1,       2,            3,  "asian",        "p75",  FALSE,   value,
                     2,       3,            4,  "black",        "p75",  FALSE,   value,
                     3,       4,            5,  "white",        "p75",  FALSE,   value)

如果有任何帮助，我将不胜感激!

Answer 1

如果我明白你想要什么，你可以做这样的事情

(首先，快速解释一下我使用的 pivot_longer() 选项)

• cols 选择您要旋转的列
• names_to 具有多个元素，我们可以将其转换为多个列
• names_prefix 允许我们删除要转换的列之间的公共(public)前缀
• name_sep 允许我们指示 names_to 中的多个名称在原始列中是如何分解的

pivot_longer(
  df,
  cols = starts_with("kfr"),
  names_to=c("race", "pooled", "percentile", "se"), 
  names_prefix="kfr_", 
  names_sep="_"
) %>% 
  select(!pooled) %>% 
  mutate(se=!is.na(se))


# A tibble: 14 x 7
   state county kids_count race   percentile se    value
   <dbl>  <dbl>      <dbl> <chr>  <chr>      <lgl> <dbl>
 1     1      2         11 asian  p25        FALSE     3
 2     1      2         11 asian  p75        FALSE     4
 3     1      2         11 black  p25        FALSE     5
 4     1      2         11 black  p75        FALSE     6
 5     1      2         11 pooled p25        FALSE     7
 6     1      2         11 pooled p75        FALSE     8
 7     1      2         11 white  p25        FALSE     9
 8     1      2         11 white  p75        FALSE    10
 9     1      2         11 asian  p25        TRUE     12
10     1      2         11 asian  p75        TRUE     13
11     1      2         11 black  p25        TRUE     14
12     1      2         11 black  p75        TRUE     15
13     1      2         11 white  p25        TRUE     16
14     1      2         11 white  p75        TRUE     17