一、漏斗分析

先看漏斗的效果

其实就是根据用户的行为一层一层分析用户的转化率。

活跃–>商品详情页–>下单–>支付

每一个指标对应的表：
活跃：dws_user_active_history
商品详情页：dwd_good_item
下单：dwd_user_order
支付：dwd_user_order

二、思路

首先统计当天活跃用户数量。
接着统计当天进入了多少个商品详情页。
接着统计当天下单的数量。
最后统计当天支付的数量。

并且计算每一层的转化率。
最终把结果数据保存到表app_user_conver_funnel中。

三、app层

1、app_user_conver_funnel

（1）源表

SQL

（2）建表语句

create external table if not exists app_mall.app_user_conver_funnel(
    active_num    int,
    item_num     int,
    order_num    int,
    pay_num    int,
    active_to_item_ratio    decimal(10,2),
    item_to_order_ratio    decimal(10,2),
    order_to_pay_ratio    decimal(10,2)
)partitioned by(dt string) 
 row format delimited  
 fields terminated by '\t'
 location 'hdfs://bigdata01:9000/data/app/user_conver_funnel/';

（3）映射关系

insert overwrite table app_mall.app_user_conver_funnel partition(dt='20260201')  select 
duah.active_num,
dgi.item_num,
duo.order_num,
duo.pay_num,
dgi.item_num/duah.active_num as active_to_item_ratio,
duo.order_num/dgi.item_num as item_to_order_ratio,
duo.pay_num/duo.order_num as order_to_pay_ratio
from
(
    select 
	count(*) as active_num
	from dws_mall.dws_user_active_history
	where dt = '20260201'
) as duah
join
(
    select 
	count(distinct goods_id) as item_num
	from dwd_mall.dwd_good_item
    where dt = '20260201'
)as dgi
on 1=1
join
(
    select
	count(*) as order_num,
	sum(case when order_status != 0 then 1 else 0 end) as pay_num
	from dwd_mall.dwd_user_order
	where dt = '20260201'
) as duo
on 1=1;

五、开发脚本

1、表初始化脚本（初始化执行一次）

app_mall_init_table_4.sh

内容如下：

#!/bin/bash
# 需求四：漏斗分析
# app层数据库和表初始化脚本，只需要执行一次即可

hive -e "
create database if not exists app_mall;

create external table if not exists app_mall.app_user_conver_funnel(
    active_num    int,
    item_num     int,
    order_num    int,
    pay_num    int,
    active_to_item_ratio    decimal(10,2),
    item_to_order_ratio    decimal(10,2),
    order_to_pay_ratio    decimal(10,2)
)partitioned by(dt string) 
 row format delimited  
 fields terminated by '\t'
 location 'hdfs://bigdata01:9000/data/app/user_conver_funnel/';
"

2、添加分区脚本（每天执行一次）

app_mall_add_partition_4.sh

内容如下：

#!/bin/bash
# 需求四：漏斗分析
# 每天凌晨执行一次

# 默认获取昨天的日期，也支持传参指定一个日期
if [ "z$1" = "z" ]
then
dt=`date +%Y%m%d --date="1 days ago"`
else
dt=$1
fi

hive -e "
insert overwrite table app_mall.app_user_conver_funnel partition(dt='${dt}')  select 
duah.active_num,
dgi.item_num,
duo.order_num,
duo.pay_num,
dgi.item_num/duah.active_num as active_to_item_ratio,
duo.order_num/dgi.item_num as item_to_order_ratio,
duo.pay_num/duo.order_num as order_to_pay_ratio
from
(
    select 
	count(*) as active_num
	from dws_mall.dws_user_active_history
	where dt = '${dt}'
) as duah
join
(
    select 
	count(distinct goods_id) as item_num
	from dwd_mall.dwd_good_item
    where dt = '${dt}'
)as dgi
on 1=1
join
(
    select
	count(*) as order_num,
	sum(case when order_status != 0 then 1 else 0 end) as pay_num
	from dwd_mall.dwd_user_order
	where dt = '${dt}'
) as duo
on 1=1;
"

六、执行脚本

1、先执行初始化脚本

sh app_mall_init_table_4.sh

2、再执行添加分区脚本

sh  app_mall_add_partition_4.sh 20260201

这个要等一会，估计5分钟，如下就好了;

七、验证结果

select * from app_mall.app_user_conver_funnel where dt = '20260201';