时间戳

首页 > 解决方案 > 使用 dplyr 按组按最低和最高年份过滤

r - 使用 dplyr 按组按最低和最高年份过滤

问题描述

我觉得这里的答案很明显，但我无法确定。我有这个数据框：

df <- structure(list(SIC = c(3L, 12L, 11L, 7L, 18L, 5L, 19L, 17L, 1L, 
10L, 8L, 16L, 14L, 2L, 15L, 6L, 9L, 4L, 13L, 3L, 12L, 11L, 7L, 
18L, 5L, 19L, 17L, 1L, 10L, 8L, 16L, 14L, 2L, 15L, 6L, 9L, 4L, 
13L, 3L, 12L, 11L, 7L, 18L, 5L, 19L, 17L, 1L, 10L, 8L, 16L, 14L, 
2L, 15L, 6L, 9L, 4L, 13L, 3L, 12L, 11L, 7L, 18L, 5L, 19L, 17L, 
1L, 10L, 8L, 16L, 14L, 2L, 15L, 6L, 9L, 4L, 13L, 3L, 12L, 11L, 
7L, 18L, 5L, 19L, 17L, 1L, 10L, 8L, 16L, 14L, 2L, 15L, 6L, 9L, 
4L, 13L, 3L, 12L, 11L, 7L, 18L, 5L, 19L, 17L, 1L, 10L, 8L, 16L, 
14L, 2L, 15L, 6L, 9L, 4L, 13L, 3L, 12L, 11L, 7L, 18L, 5L, 19L, 
17L, 1L, 10L, 8L, 16L, 14L, 2L, 15L, 6L, 9L, 4L, 13L, 3L, 12L, 
11L, 7L, 18L, 5L, 19L, 17L, 1L, 10L, 8L, 16L, 14L, 2L, 15L, 6L, 
9L, 4L, 13L), year = c(2011, 2011, 2011, 2011, 2011, 2011, 2011, 
2011, 2011, 2011, 2011, 2011, 2011, 2011, 2011, 2011, 2011, 2011, 
2011, 2012, 2012, 2012, 2012, 2012, 2012, 2012, 2012, 2012, 2012, 
2012, 2012, 2012, 2012, 2012, 2012, 2012, 2012, 2012, 2013, 2013, 
2013, 2013, 2013, 2013, 2013, 2013, 2013, 2013, 2013, 2013, 2013, 
2013, 2013, 2013, 2013, 2013, 2013, 2014, 2014, 2014, 2014, 2014, 
2014, 2014, 2014, 2014, 2014, 2014, 2014, 2014, 2014, 2014, 2014, 
2014, 2014, 2014, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 
2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 2015, 
2016, 2016, 2016, 2016, 2016, 2016, 2016, 2016, 2016, 2016, 2016, 
2016, 2016, 2016, 2016, 2016, 2016, 2016, 2016, 2017, 2017, 2017, 
2017, 2017, 2017, 2017, 2017, 2017, 2017, 2017, 2017, 2017, 2017, 
2017, 2017, 2017, 2017, 2017, 2018, 2018, 2018, 2018, 2018, 2018, 
2018, 2018, 2018, 2018, 2018, 2018, 2018, 2018, 2018, 2018, 2018, 
2018, 2018), value = c(NA, 0.081, 0.218, 0.212, NA, 0.092, 0.142, 
0.001, 0.045, 0.143, 0.361, 0.175, 0.295, 0.003, 0.146, 0.01, 
0.163, NA, 0.225, NA, 0.108, 0.274, 0.219, NA, 0.097, 0.148, 
-0.049, 0.098, 0.26, 0.251, 0.153, 0.262, 0.119, 0.096, 0, 0.149, 
NA, NA, NA, 0.064, 0.27, 0.16, NA, 0.103, 0.148, -0.029, 0.084, 
0.219, 0.314, 0.142, 0.255, 0.026, 0.031, -0.003, 0.164, NA, 
NA, NA, NA, 0.257394804, 0.124025397, NA, 0.071727544, 0.13439, 
-0.017419321, 0.091993981, 0.171021874, 0.308369685, 0.208573024, 
0.310316421, 0.06216349, 0.074913633, -0.034273066, 0.181129287, 
0.07876301, 0.121, NA, -0.063226494, 0.233968039, 0.179367136, 
NA, 0.105362761, 0.15319907, -0.033967241, -0.035027867, 0.144316565, 
0.304955404, 0.069662044, 0.304262651, 0.075256422, 0.051273353, 
-0.067541918, 0.157814304, 0.050231459, 0.06308377, NA, -8.4, 
21, 17.9, NA, 7.3, 12.6, -1.2, 4.1, 10.3, 30.2, 8.7, 28.1, 4.7, 
1.3, -7.7, 12.5, 15.9, 19.4, 16.9, 4, 18.2, 13.5, NA, 10.9, 12.8, 
-0.7, 4.2, 7.5, 26.8, 5, 30.3, 0.9, 2.5, -2.4, 13.5, 12.8, 17, 
NA, NA, 17.2, 17.7, NA, 0.6, 11.6, -2.9, 3, 18.7, 31, 6.2, 30.1, 
-1.1, 5.7, -0.5, 13.6, 6.1, -7)), class = c("tbl_df", "tbl", 
"data.frame"), row.names = c(NA, -152L))

我想按最低和最高年份过滤它。这样所有值都有最低和最高年份列。我将split/spread并比较这两列。我的方法是先group_by在 SIC 上，然后filter在 上!is.na(value)，但这会返回很少的值。有一组 SIC 只有一个值，因此其最低年份和最高年份应该相同。这是我迄今为止尝试过的并且过滤了太多的值。

library(dplyr)
df %>% 
  group_by(SIC) %>% 
  filter(!is.na(value)) %>% 
  filter(year %in% c(min(year), max(year)))

# A tibble: 35 x 3
# Groups:   SIC [18]
     SIC  year value
   <int> <dbl> <dbl>
 1    12  2011 0.081
 2    11  2011 0.218
 3     7  2011 0.212
 4     5  2011 0.092
 5    19  2011 0.142
 6    17  2011 0.001
 7     1  2011 0.045
 8    10  2011 0.143
 9     8  2011 0.361
10    16  2011 0.175
# ... with 25 more rows

有任何想法吗？谢谢。

编辑：

数据的一个简单版本是像这样过滤：

tibble(
  SIC = c(1,1,1,2,2, 2), 
  year = c(2011, 2012, 2013, 2011, 2012, 2013), 
  value = c(3, 4, NA, NA, 4, NA)
) %>% 
  filter(!is.na(value)) 

# A tibble: 3 x 3
    SIC  year value
  <dbl> <dbl> <dbl>
1     1  2011     3
2     1  2012     4
3     2  2012     4

所有出现一次的行都应计为最小值和最大值。有没有办法为那些在应用过滤器后只出现一次的行创建重复行？

标签： rfilterdplyr

解决方案

我的理解是，您希望 SIC 有两行，其最大年份和最小年份相同。我认为您可以将两者拆分并绑定，以便它仍然有 2 行。

library(dplyr)

min_year <- df %>% 
  group_by(SIC) %>% 
  filter(!is.na(value)) %>% 
  filter(year %in% c(min(year)))

max_year <-  df %>% 
  group_by(SIC) %>% 
  filter(!is.na(value)) %>% 
  filter(year %in% c(max(year)))

total <- min_year %>% rbind(max_year)

推荐阅读