Question

我在names_to参数上看到了另外两个类似的问题。但是，我仍然不确定我是否会使用它。

考虑以下df：

df <- data.frame("Metric" = c("a_exp_2001","a_inc_2001","a_inc_2002","a_exp_2002"),
                "John" = c(220,230,240,250),
                "Abby" = c(440,450,470,480))

head(df)

      Metric John Abby
1 a_exp_2001  220  440
2 a_inc_2001  230  450
3 a_inc_2002  240  470
4 a_exp_2002  250  480

我正在寻找一个新的数据框架，该框架仅具有一年（例如2002年）的上述信息，但格式较长。也就是说，我希望以3列“名称”，“支出”（值对应于a_exp_2002），“收入”（值对应a_inc_2002）结束。

我不确定如何在names_to中指定pivot_longer()参数来获得最终的df。

到目前为止，我只能做到这一点：

df %>% 
   pivot_longer(cols = -c(Metric), ...)

Answer 1

如果我正确理解了您的问题，这可能是解决方案：

library(dplyr)
library(tidyr)

df <- data.frame("Metric" = c("a_exp_2001","a_inc_2001","a_inc_2002","a_exp_2002"),
                 "John" = c(220,230,240,250),
                 "Abby" = c(440,450,470,480))

df %>% 
  # get relevant info from Metric in two columns
  dplyr::mutate(IN_or_EX =  substr(Metric, start = 3, stop = 5),
                YEAR = substr(Metric, start = 7, stop = 10)) %>% 
  # filter for the year you wan
  dplyr::filter(YEAR == 2001) %>% 
  # drop the Metric column as it presents no value anymore
  dplyr::select(-Metric) %>%
  # make it longer
  tidyr::pivot_longer(-c(IN_or_EX, YEAR), names_to = "NAME", values_to = "VALUES") %>% 
  # make it wider
  tidyr::pivot_wider(names_from = IN_or_EX, values_from = VALUES, values_fill = 0)


  YEAR  NAME    exp   inc
  <chr> <chr> <dbl> <dbl>
1 2001  Abby    440   450
2 2001  John    220   230

Answer 2

您可能正在寻找：

pivot_longer(df, -1) %>% 
  separate(Metric, sep = 6, into = c("type", "year")) %>% 
  pivot_wider(names_from = type, values_from = value) %>%
  rename(Expenditure = a_exp_, Income = a_inc_)
#> # A tibble: 4 x 4
#>   year  name  Expenditure Income
#>   <chr> <chr>       <dbl>  <dbl>
#> 1 2001  John          220    230
#> 2 2001  Abby          440    450
#> 3 2002  John          250    240
#> 4 2002  Abby          480    470

Answer 3

这是工作吗？

library(tidyverse)

data <- tibble::tribble(
         ~Metric.John.Abby,
  "1 a_exp_2001  220  440",
  "2 a_inc_2001  230  450",
  "3 a_inc_2002  240  470",
  "4 a_exp_2002  250  480"
  ) %>% 
  tidyr::separate("Metric.John.Abby", into = c('num', 'metric', 'sp', 'john', 'sp2', 'abby'), sep = " ") %>% 
  select(metric, john, abby) 

data %>% 
  filter(str_detect(metric, "2002")) %>% 
  pivot_longer(cols = -metric, names_to = "names") %>% 
  separate(metric, into = c('letter', 'type', 'year')) %>% 
  select(type, names, value) %>% 
  pivot_wider(type, id_cols = c('type', 'names')) %>% 
  relocate(exp, .before = inc)
#> # A tibble: 2 x 3
#>   names exp   inc  
#>   <chr> <chr> <chr>
#> 1 john  250   240  
#> 2 abby  480   470

^{由reprex package（v0.3.0）于2020-11-08创建}

如何在pivot_longer（）中使用names_to参数

3 个答案: