在dplyr中应用过滤器时，保持group_by的机智

我试图在一个变种中应用一个过滤器，但我还没有找到正确的方法来应用过滤器，同时保持数据帧分组的灵活性

这里有一个简单的可复制示例：

# Sample data
my_dates = seq(as.Date("2020/1/1"), by = "month", length.out = 6) 
grp      = c(rep("A",3), rep("B", 3))
x        = c(2,4,6,8,10,12)

my_df <- data.frame(my_dates, grp, x)

    my_dates grp  x
1 2020-01-01   A  2
2 2020-02-01   A  4
3 2020-03-01   A  6
4 2020-04-01   B  8
5 2020-05-01   B 10
6 2020-06-01   B 12


# Pick a max date for which the data will be filtered
max_date <- "2020-05-01"


# Try to get the average by group, after filtering out the max date included
filt_data  <- my_df %>% 
  group_by(grp) %>% 
  mutate(included_data = my_dates < max_date,
         my_mean       = mean(filter(., my_dates < max_date)$x)
  )


# A tibble: 6 x 5
# Groups:   grp [2]
  my_dates   grp       x included_data my_mean
  <date>     <fct> <dbl> <lgl>           <dbl>
1 2020-01-01 A         2 TRUE                5
2 2020-02-01 A         4 TRUE                5
3 2020-03-01 A         6 TRUE                5
4 2020-04-01 B         8 TRUE                5
5 2020-05-01 B        10 FALSE               5
6 2020-06-01 B        12 FALSE               5

  my_dates   grp       x included_data my_mean
  <date>     <fct> <dbl> <lgl>           <dbl>
1 2020-01-01 A         2 TRUE                4
2 2020-02-01 A         4 TRUE                4
3 2020-03-01 A         6 TRUE                4
4 2020-04-01 B         8 TRUE                8
5 2020-05-01 B        10 FALSE               8
6 2020-06-01 B        12 FALSE               8

#示例数据
my_dates=序号（截止日期（“2020/1/1”），by=“month”，length.out=6）
grp=c（代表（“A”，3），代表（“B”，3））
x=c（2,4,6,8,10,12）
my_df在这里，最好使用“included_data”中的索引对“x”列进行子集化，而不是执行另一个
筛选

library(dplyr)
my_df %>% 
     group_by(grp) %>%
     mutate(included_data = my_dates < max_date, 
             my_mean = mean(x[included_data])) %>%
     ungroup

啊，谢谢！一个小注释，你可以删除你的一个“平均值”：
my_mean=mean（x[包含的数据]）
# A tibble: 6 x 5
#  my_dates   grp       x included_data my_mean
#  <date>     <chr> <dbl> <lgl>           <dbl>
#1 2020-01-01 A         2 TRUE                4
#2 2020-02-01 A         4 TRUE                4
#3 2020-03-01 A         6 TRUE                4
#4 2020-04-01 B         8 TRUE                8
#5 2020-05-01 B        10 FALSE               8
#6 2020-06-01 B        12 FALSE               8

my_df %>%
    group_by(grp) %>%
    mutate(included_data = my_dates < max_date, 
    my_mean = mean(filter(cur_data(), my_dates < max_date)$x)) %>% 
    ungroup
# A tibble: 6 x 5
#  my_dates   grp       x included_data my_mean
#  <date>     <chr> <dbl> <lgl>           <dbl>
#1 2020-01-01 A         2 TRUE                4
#2 2020-02-01 A         4 TRUE                4
#3 2020-03-01 A         6 TRUE                4
#4 2020-04-01 B         8 TRUE                8
#5 2020-05-01 B        10 FALSE               8
#6 2020-06-01 B        12 FALSE               8