Hive的CTE 公共表达式

目录

1.语法

[2. 使用场景](#2. 使用场景)

select语句

[chaining CTEs 链式](#chaining CTEs 链式)

union语句

[insert into 语句](#insert into 语句)

[create table as 语句](#create table as 语句)

前言

Common Table Expressions(CTE):公共表达式是一个临时的结果集,该结果集是从with子句中指定的查询派生而来的,紧跟在select 或 insert关键字之前。CTE可以在 select,insert, create table as select 等语句中使用。

1.语法

sql 复制代码
[wtih CommonTableExpression]
select
        column1,
        column2, ...
from table 
[where 条件] 
[group by column]
[order by column] 
[cluster by column| [distribute by column] [sort by column] 
[limit [offset,] rows];

2. 使用场景

select语句

sql 复制代码
with tmp as (
    select
        oid,
        uid,
        otime,
        date_format(otime, 'yyyy-MM') as dt,
        oamount,
        ---计算rk的目的是为了获取记录中的第一条
        row_number() over (partition by uid,date_format(otime, 'yyyy-MM') order by otime) rk
    from t_order
)
 select
    uid,
    --每个用户一月份的订单数
    sum(if(dt = '2018-01', 1, 0)) as  m1_count,
    --每个用户二月份的订单数
    sum(if(dt = '2018-02', 1, 0)) as  m2_count
from tmp
 group by uid
 having m1_count >0 and m2_count=0;

chaining CTEs 链式

sql 复制代码
with tmp1 as (
    select
        oid,
        uid,
        otime,
        date_format(otime, 'yyyy-MM') as dt,
        oamount,
        ---计算rk的目的是为了获取记录中的第一条
        row_number() over (partition by uid,date_format(otime, 'yyyy-MM') order by otime) as rk
    from t_order
),
     tmp2 as
         (select
              uid,
              --每个用户一月份的订单数
              sum(if(dt = '2018-01', 1, 0)) as m1_count,
              --每个用户二月份的订单数
              sum(if(dt = '2018-02', 1, 0)) as m2_count
          from tmp1
          group by uid
          having m1_count > 0
             and m2_count = 0)
select * from tmp2 limit 1;

union语句

sql 复制代码
with q1 as (select * from student where num = 95002),
     q2 as (select * from student where num = 95004)
select * from q1 union all select * from q2;

insert into 语句

sql 复制代码
with tmp1 as (
    select
        oid,
        uid,
        otime,
        date_format(otime, 'yyyy-MM') as dt,
        oamount,
        ---计算rk的目的是为了获取记录中的第一条
        row_number() over (partition by uid,date_format(otime, 'yyyy-MM') order by otime) as rk
    from t_order
),
     tmp2 as
         (select
              uid,
              --每个用户一月份的订单数
              sum(if(dt = '2018-01', 1, 0)) as m1_count,
              --每个用户二月份的订单数
              sum(if(dt = '2018-02', 1, 0)) as m2_count
          from tmp1
          group by uid
          having m1_count > 0
             and m2_count = 0)

insert into tmp3
select * from tmp2 limit 10;

create table as 语句

sql 复制代码
--- 从tmp2 表中取10条数据,基于此创建表tmp3 
create table tmp3 as 
with tmp1 as (
    select
        oid,
        uid,
        otime,
        date_format(otime, 'yyyy-MM') as dt,
        oamount,
        ---计算rk的目的是为了获取记录中的第一条
        row_number() over (partition by uid,date_format(otime, 'yyyy-MM') order by otime) as rk
    from t_order
),
     tmp2 as
         (select
              uid,
              --每个用户一月份的订单数
              sum(if(dt = '2018-01', 1, 0)) as m1_count,
              --每个用户二月份的订单数
              sum(if(dt = '2018-02', 1, 0)) as m2_count
          from tmp1
          group by uid
          having m1_count > 0
             and m2_count = 0)
select * from tmp2 limit 10;
相关推荐
chat2tomorrow8 分钟前
数据采集平台的起源与演进:从ETL到数据复制
大数据·数据库·数据仓库·mysql·低代码·postgresql·etl
DashingGuy1 天前
数仓建模理论
数据仓库
像豆芽一样优秀1 天前
Hive和Flink数据倾斜问题
大数据·数据仓库·hive·hadoop·flink
howard20052 天前
VMWare上搭建Hive集群
hive·hadoop
aristo_boyunv2 天前
拦截器和过滤器(理论+实操)
java·数据仓库·hadoop·servlet
元媛媛2 天前
数据仓库概要
数据仓库
cg.family2 天前
Doris 数据仓库例子
数据仓库·doris
TDengine (老段)2 天前
从 ETL 到 Agentic AI:工业数据管理变革与 TDengine IDMP 的治理之道
数据库·数据仓库·人工智能·物联网·时序数据库·etl·tdengine
镜舟科技3 天前
告别 Hadoop,拥抱 StarRocks!政采云数据平台升级之路
大数据·starrocks·数据仓库·hadoop·存算分离