pyspark做透视表

了解到  pivot()函数可以实现

尝试一：

df = spark.sql("""
select payload.avatar_id as avatar,
       local_dt as play_date,
       count(distinct payload.account_id) as player_count
from dts_obt.eventtypeingameitems 
where dt between "20200501" and "20200504"
  and local_dt between "20200502" and "20200503"
group by payload.avatar_id,local_dt
order by payload.avatar_id,local_dt
""")
df.show()

　　

 这里我想要（以下是通过excel表格的透视表实现的）

 如何用pyspark实现？

df.pivot(index="avatar",columns="play_date",values="player_count")

　但是报错