思路是这样的：
把nginx日志数据清洗后导入avro表，创建分区表etl_nginx_avro，以month（202005）/day(01)进行分区，创建中间表tmp_etl_nginx_avro不分区，中间表比分区表多了两个字段，month和hour（其实是day创建时没留神）；
中间表tmp_etl_nginx_avro：
分区表etl_nginx_avro：
自动分区：
set hive.exec.dynamic.partition=true;
set hive.exec.dynamic.partition.mode=nonstrict;
insert overwrite table etl_nginx_avro partition(month,day)
select ip,acctime,locations,status,request_body,referer,useragent,osfamily,osname,uafamily,devicetype,month,hour from tmp_etl_nginx_avro;
然后分区字段：
但是分区文件夹变成这样：
中间表数据正常：
>select ip,acctime,locations,status,request_body,referer,useragent,osfamily,osname,uafamily,devicetype,month,hour from tmp_etl_nginx_avro limit 10;
182.254.52.17 20200503 /topics/809 200 - http://hainiubl.com/topics/809 Browser Windows Windows 7 Firefox unknown 202005 03
182.254.52.17 20200503 /topics/124?page%3D1 200 - http://hainiubl.com/topics/124?page%3D1 Browser Windows Windows 7 Firefox unknown 202005 03
自动分区完后 分区表数据正常：
> select * from etl_nginx_avro limit 10;
OK
123.4.226.100 20200501 /topics/44 200 - http://www.hainiubl.com/ Browser Windows Windows Chrome unknown 202005 01
123.4.226.100 20200501 /topics/181 301 - http://www.hainiubl.com/ Browser Windows Windows Chrome unknown 202005 01
123.4.226.100 20200501 /topics/181 200 - http://www.hainiubl.com/ Browser Windows Windows Chrome unknown 202005 01
123.4.226.100 20200501 /topics/182 301 - http://www.hainiubl.com/ Browser Windows Windows Chrome unknown 202005 01
123.4.226.100 20200501 /topics/182 200 - http://www.hainiubl.com/ Browser Windows Windows Chrome unknown 202005 01
就是这个分区名字有点辣眼睛，百思不得其解
