为什么dplyr 0.8.3仍然使用lag运算符和分组数据生成NA?

我想用dplyr延迟分组数据中的变量。我用 lag dplyr lag function returns NAs 有人指出 https://github.com/tidyverse/dplyr/issues/1540 哈雷在2016年修正了一些错误。所以,我想已经解决了。为什么我的延迟命令仍然抛出NA?

library(tidyverse)

data = data.frame(id=c(1,1,1,2,2,2,3,3,3,4,4,4), time=seq(1:3), x=rep(c(5:8), each=3))

data %>%
  group_by(id) %>%
  mutate(x_lag = lag(x, n=1, default=NA, order_by=TRUE)) %>% 
  select(id, time, x, x_lag) 

data %>%
  group_by(id) %>%
  mutate(x_lag = lag(x, n=1, default=NA, order_by=FALSE)) %>% 
  select(id, time, x, x_lag) 

data %>%
  group_by(id) %>%
  arrange(id) %>%
  mutate(x_lag = lag(x, n=1, default=NA, order_by=FALSE)) %>% 
  select(id, time, x, x_lag) 

data %>%
  group_by(id) %>%
  mutate(x_lag = lag(x, n=1, default=0, order_by=TRUE)) %>% 
  select(id, time, x, x_lag) 

# A tibble: 8 x 4
# Groups:   id, time [12]
      id  time     x x_lag
   <dbl> <int> <int> <int>
 1     1     1     5    NA
 2     1     2     5    NA
 3     1     3     5    NA
 4     2     1     6    NA
 5     2     2     6    NA
 6     2     3     6    NA
 7     3     1     7    NA
 8     3     2     7    NA