在R中的多列上过滤数据，至少匹配6次

Question

我有以下数据，我只想保留数据集中只有6个实例的情况(相同的姓氏和名字)。例如，Quincy在df中出现了6次，我希望保留这些情况中的每一种，但是去掉Abrines，因为这个个体只有3个实例(< 6)。

   last  first start_year end_year Team     GP   MIN   PTS     W     L
   <chr> <chr>      <int>    <int> <chr> <int> <dbl> <dbl> <int> <int>
 1 Abri… Alex        2016     2017 OKC      68  15.5   6      37    31
 2 Abri… Alex        2017     2018 OKC      75  15.1   4.8    42    33
 3 Abri… Alex        2018     2019 OKC      31  19     5.3    21    10
 4 Acy   Quin…       2013     2014 SAC      63  13.5   2.7    22    41
 5 Acy   Quin…       2014     2015 NYK      68  18.9   5.9    12    56
 6 Acy   Quin…       2015     2016 SAC      59  14.8   5.3    21    38
 7 Acy   Quin…       2016     2017 BKN      38  14.7   5.8    11    27
 8 Acy   Quin…       2017     2018 BKN      70  19.4   5.9    26    44
 9 Acy   Quin…       2018     2019 PHX      10  12.3   1.7     2     8

我尝试过x <- df %>% count(last, first) %>% filter(n == 6)，然后是df %>% filter(last %in% x$last & first %in% x$first)，但是它分别匹配任何姓氏和任何名字，而不是匹配姓氏和名字。我相信还有一个更简单的解决方案，不需要首先使用group_by。

我希望这个解决方案看起来像：

  <chr> <chr>      <int>    <int> <chr> <int> <dbl> <dbl> <int> <int>
1 Acy   Quin…       2013     2014 SAC      63  13.5   2.7    22    41
2 Acy   Quin…       2014     2015 NYK      68  18.9   5.9    12    56
3 Acy   Quin…       2015     2016 SAC      59  14.8   5.3    21    38
4 Acy   Quin…       2016     2017 BKN      38  14.7   5.8    11    27
5 Acy   Quin…       2017     2018 BKN      70  19.4   5.9    26    44
6 Acy   Quin…       2018     2019 PHX      10  12.3   1.7     2     8
7 Adams Stev…       2013     2014 OKC      81  14.8   3.3    59    22
8 Adams Stev…       2014     2015 OKC      70  25.3   7.7    37    33
9 Adams Stev…       2015     2016 OKC      80  25.2   8      54    26
10 Adams Stev…       2016     2017 OKC      80  29.9  11.3    47    33
11 Adams Stev…       2017     2018 OKC      76  32.7  13.9    43    33
12 Adams Stev…       2018     2019 OKC      80  33.4  13.8    47    33

Answer 1

不需要count summarise数据，创建一个新对象，然后执行filter，我们可以根据条件直接对组进行group_by、‘data’、'first‘。

library(dplyr)
df1 <- df %>%
         group_by(last, first) %>%
         filter(n() == 6)

如果至少为6，则更改==或>=

或者另一个选择是table

subset(df,   paste(last, first) %in% names(which(table(paste(last, first)) == 6)))

Answer 2

在基本R中，我们可以使用ave来计算每组first和last值中的行数，并选择行数为6的组。

subset(df, ave(start_year, first, last, FUN = length) == 6)

#  last first start_year end_year Team GP  MIN PTS  W  L
#4  Acy Quin…       2013     2014  SAC 63 13.5 2.7 22 41
#5  Acy Quin…       2014     2015  NYK 68 18.9 5.9 12 56
#6  Acy Quin…       2015     2016  SAC 59 14.8 5.3 21 38
#7  Acy Quin…       2016     2017  BKN 38 14.7 5.8 11 27
#8  Acy Quin…       2017     2018  BKN 70 19.4 5.9 26 44
#9  Acy Quin…       2018     2019  PHX 10 12.3 1.7  2  8

我们也可以用data.table做同样的事情

library(data.table)
setDT(df)[,.SD[.N == 6], .(first, last)]

数据

df <- structure(list(last = structure(c(1L, 1L, 1L, 2L, 2L, 2L, 2L, 
2L, 2L), .Label = c("Abri…", "Acy"), class = "factor"), first = structure(c(1L, 
1L, 1L, 2L, 2L, 2L, 2L, 2L, 2L), .Label = c("Alex", "Quin…"
), class = "factor"), start_year = c(2016L, 2017L, 2018L, 2013L, 
2014L, 2015L, 2016L, 2017L, 2018L), end_year = c(2017L, 2018L, 
2019L, 2014L, 2015L, 2016L, 2017L, 2018L, 2019L), Team = structure(c(3L, 
3L, 3L, 5L, 2L, 5L, 1L, 1L, 4L), .Label = c("BKN", "NYK", "OKC", 
"PHX", "SAC"), class = "factor"), GP = c(68L, 75L, 31L, 63L, 
68L, 59L, 38L, 70L, 10L), MIN = c(15.5, 15.1, 19, 13.5, 18.9, 
14.8, 14.7, 19.4, 12.3), PTS = c(6, 4.8, 5.3, 2.7, 5.9, 5.3, 
5.8, 5.9, 1.7), W = c(37L, 42L, 21L, 22L, 12L, 21L, 11L, 26L, 
2L), L = c(31L, 33L, 10L, 41L, 56L, 38L, 27L, 44L, 8L)), class = "data.frame", 
row.names = c(NA, -9L))