SQL统计连续登陆3天的用户(连续活跃超3天用户)

目录

1. 数据准备

-- 数据准备
WITH user_active_info AS (
SELECT * FROM (
    VALUES ('10001' , '2023-02-01'),('10001' , '2023-02-03')
          ,('10001' , '2023-02-04'),('10001' , '2023-02-05')
          ,('10002' , '2023-02-02'),('10002' , '2023-02-03')
          ,('10002' , '2023-02-04'),('10002' , '2023-02-05')
          ,('10002' , '2023-02-07'),('10003' , '2023-02-02')
          ,('10003' , '2023-02-03'),('10003' , '2023-02-04')
          ,('10003' , '2023-02-05'),('10003' , '2023-02-06')
          ,('10003' , '2023-02-07'),('10003' , '2023-02-08')
          ,('10004' , '2023-02-03'),('10004' , '2023-02-04')
          ,('10004' , '2023-02-06'),('10004' , '2023-02-07')
          ,('10004' , '2023-02-08'),('10004' , '2023-02-08') 
    	  ,('10005' , '2023-02-02'),('10005' , '2023-02-05') 
) AS user_active_info(user_id, active_date) 
)

2. 方法一: 差值计算

-- 1. 对用户数据进行分组,按照活跃日期进行排序(去重:防止有一天有多次活跃记录)
SELECT 
      user_id
    , active_date
    , ROW_NUMBER() OVER(PARTITION BY user_id ORDER BY active_date) AS rn 
FROM user_active_info
GROUP BY user_id , active_date
;
user_idactive_datern
100012023-02-011
100012023-02-032
100012023-02-043
100012023-02-054
100022023-02-021
100022023-02-032
100022023-02-043
100022023-02-054
100022023-02-075
………
-- 2. 使用活跃日期和排序rn进行差值计算,得到的日期如果是相等的,就说明活跃日期是连续的
SELECT 
      user_id
    , active_date
    , rn 
    , DATE_SUB(active_date,rn) AS sub_date
FROM (
    SELECT 
          user_id
        , active_date
        , ROW_NUMBER() OVER(PARTITION BY user_id ORDER BY active_date) AS rn 
    FROM user_active_info 
    GROUP BY user_id , active_date
    ) a
;
user_idactive_daternsub_date
100012023-02-0112023-01-31
100012023-02-0322023-02-01
100012023-02-0432023-02-01
100012023-02-0542023-02-01
100022023-02-0212023-02-01
100022023-02-0322023-02-01
100022023-02-0432023-02-01
100022023-02-0542023-02-01
100022023-02-0752023-02-02
…………
-- 3. 按照user_id和sub_date 进行分组求和,筛选出连续登陆天数大于3天的用户
SELECT 
      user_id
    , MIN(active_date) AS begin_date
    , MAX(active_date) AS end_date
    , COUNT (1) AS login_duration
FROM (
    SELECT 
          user_id
        , active_date
        , rn 
        , DATE_SUB(active_date,rn) AS sub_date
    FROM (
        SELECT 
              user_id
            , active_date
            , ROW_NUMBER() OVER(PARTITION BY user_id ORDER BY active_date) AS rn 
        FROM user_active_info 
        GROUP BY user_id , active_date
    ) a
) b
GROUP BY user_id , sub_date
HAVING login_duration >= 3
;
user_idbegin_dateend_datelogin_duration
100012023-02-032023-02-053
100022023-02-022023-02-054
100032023-02-022023-02-087
100042023-02-062023-02-083

3. 方法二: lead或lag函数

-- 1. 将active_date 上抬2行,不存在默认为'0'(计算连续活跃3天以上的, 上抬2行,n天上抬n-1行)(去重:防止有一天有多次活跃记录)
SELECT 
      user_id
    , active_date
    , lead(active_date , 2 , 0) OVER(PARTITION BY user_id ORDER BY active_date) AS lead_active_date 
FROM user_active_info
GROUP BY user_id , active_date
user_idactive_datelead_active_date
100012023-02-012023-02-04
100012023-02-032023-02-05
100012023-02-040
100012023-02-050
100022023-02-022023-02-04
100022023-02-032023-02-05
100022023-02-042023-02-07
100022023-02-050
100022023-02-070
………
-- 2. 过滤筛选出, lead_active_date 与 active_date 差值为2的, 差值2 -> 连续活跃了3天
SELECT 
      user_id , active_date , lead_active_date
FROM (
    SELECT 
          user_id
        , active_date
        , lead(active_date , 2 , 0) OVER(PARTITION BY user_id ORDER BY active_date) AS lead_active_date
    FROM user_active_info
    GROUP BY user_id , active_date
) a 
WHERE  lead_active_date != '0'
    AND DATEDIFF(lead_active_date , active_date) = 2
user_idactive_datelead_active_date
100012023-02-032023-02-05
100022023-02-022023-02-04
100022023-02-032023-02-05
………
-- 3. user_id 去重, 得到连续活跃天数>=3天的用户
SELECT 
      user_id
FROM (
    SELECT 
          user_id , active_date , lead_active_date
    FROM (
        SELECT 
              user_id
            , active_date
            , lead(active_date , 2 , 0) OVER(PARTITION BY user_id ORDER BY active_date) AS lead_active_date 
        FROM user_active_info
        GROUP BY user_id , active_date
    ) a 
    WHERE  lead_active_date != '0'
        AND DATEDIFF(lead_active_date , active_date) = 2
) b
GROUP BY user_id
user_id
10001
10002
10003
10004
end
# sql # 大数据
Logo
GitCode 开源社区

旨在为数千万中国开发者提供一个无缝且高效的云端环境，以支持学习、使用和贡献开源项目。

更多推荐

[转载]在Windows环境下安装GNU Radio

转自：在Windows环境下安装GNURadio_恐弱智_新浪博客GNU Radio是用Python开发的，大部分开源的工程能够在Linux环境下运行良好，而Windows下却运行的很勉强，而且安装配置都很复杂。GNU Radio算是个例外了，不光提供了Windows的二进制安装，还有比较详细的说明。我是Python小白，所以折腾了好久才弄好，特意记录下来，免得以后再装还折腾。GNU Radio的

avatar GitCode 开源社区

centOS 8 使用dnf安装Docker

DNF是什么？CentOS 8使用YUM软件包管理器版本v4.0.4。现在，该版本使用DNF(已删除YUM)。DNF是软件包管理器。它会在Linux发行版上安装，执行更新并删除软件包。使用DNF安装Docker跳过具有损坏依赖性的程序包一个有效的解决方案是使您的CentOS 8系统使用以下--nobest命令安装最符合条件的版本：sudo dnf install docker...

avatar GitCode 开源社区

定时同步数据库表(mysql+linux+crontab)

sync.sh里面的参数需要改变，ip/username/password/database/tablesync.sh#!/bin/sh# Please change the IP and password of the data source db.# Then change the table name.filename=/home/nington/db/$(date +%Y-%m

avatar GitCode 开源社区