'SparkSession' object has no attribute '_gateway'(待解决)

背景：
源码

def save_to_db(df,table,flag=0):
    
    if flag==0:
        df.write.jdbc(url=url, table=table, properties={'driver':driver,"user": user, "password": password})
    else:
        df.write.jdbc(url=xieurl, table=table, mode='append', properties={'driver':driver,"user": user, "password": password})
    return 
  ……………………………………
 save_to_db(df=daysdf, table='t_exer_days_copy', flag=flag)

运行上一个程序正常

现在报错显示

/opt/python3.6.7/lib/python3.6/site-packages/pyspark/sql/readwriter.py in jdbc(self, url, table, mode, properties)
    981         if properties is None:
    982             properties = dict()
--> 983         jprop = JavaClass("java.util.Properties", self._spark._sc._gateway._gateway_client)()
    984         for k in properties:
    985             jprop.setProperty(k, properties[k])

AttributeError: 'SparkSession' object has no attribute '_gateway'

检查了好久了。也是心力交瘁呀！

……………………分割线
放弃观察，改变策略

def save_to_db(df,table,flag=0):
    if flag==0:
        mode='overwrite'
    else:
        mode='append'
    df.write.mode("overwrite").format("jdbc").options(
    url=xieurl,
    user=user,
    password=password,
    dbtable=table,
    batchsize="1000",
).save()