我想在spark框架中操纵日期。但我不能。查看我的代码：

library(sparklyr)
library(dplyr)
library(lubridate)

my_con <- spark_connect(master = "local", version = "3.0.0")

df <- structure(list(x = c("19991123", NA, "19980101")), class = "data.frame", row.names = c(NA, -3L))

df_1 <- sdf_copy_to(sc = my_con, x = df)

class(df_1)

# [1] "tbl_spark" "tbl_sql"   "tbl_lazy"  "tbl"

df_1 %>% 
  mutate(.data = ., my_date = as_date(x))

但是：

x        my_date   
  <chr>    <date>    
 1 19991123 NA        
 2 NA       NA        
 3 19980101 NA
``` `NA` 产生了一种新的方法。发生了什么？