1,ck对窗口函数支持不是太好,只是有限的支持。
表结构如下:
CREATE TABLE warehouse.c_click_common_16_cls (`id` String, `app_version` Nullable(String), `guid` Nullable(String), `imei` Nullable(String), `mac` Nullable(String), `channel_id` Nullable(String), `tel_opr` Nullable(String), `phone_brand` Nullable(String), `phone_type` Nullable(String), `osversion` Nullable(String), `platform` Int32, `lang` Nullable(String), `session_id` Nullable(String), `client_time` Int64, `user_id` Nullable(String), `event_id` Nullable(String), `client_ip` Nullable(String), `net_type` Int32, `body_title` Nullable(String), `body_id` Nullable(String), `body_type` Nullable(String), `body_car_id` Nullable(String), `body_city_name` Nullable(String), `app_date` Int32, `create_time` Date, `body_index` Nullable(String)) ENGINE = Distributed(crm_4shards_1replicas, warehouse, c_click_common_16, rand())
2,业务数据分行:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25select versiontype,platform,date, arrayJoin(arr_val) --arrayJoin(row_number) from ( select versiontype,platform,date, groupArray(4)(uv) AS arr_val, arrayEnumerate(4)(arr_val) as row_number from ( select case when app_version>='3.0.0' then 'new' else 'old' end as versiontype, platform, toDate(client_time/1000) date, body_title, count(distinct imei) uv from warehouse.c_click_common_16_cls where toDate(client_time/1000)='2020-06-22' and body_title not in('xxxx') group by versiontype, platform, toDate(client_time/1000), body_title order by uv desc) m group by versiontype,platform,date ) c
查询结果如下:
3,针对结果数据进行分行
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25select versiontype,platform,date, arrayJoin(arr_val) --arrayJoin(row_number) from ( select versiontype,platform,date, groupArray(4)(uv) AS arr_val, arrayEnumerate(4)(arr_val) as row_number from ( select case when app_version>='3.0.0' then 'new' else 'old' end as versiontype, platform, toDate(client_time/1000) date, body_title, count(distinct imei) uv from warehouse.c_click_common_16_cls where toDate(client_time/1000)='2020-06-22' and body_title not in('xxxx') group by versiontype, platform, toDate(client_time/1000), body_title order by uv desc) m group by versiontype,platform,date ) c
查询结果如下:
最后
以上就是完美芝麻最近收集整理的关于ClickHouse支持row_number函数,相关的demo如下的全部内容,更多相关ClickHouse支持row_number函数,相关内容请搜索靠谱客的其他文章。
发表评论 取消回复