查询过去一年的活跃订阅人数和订阅人数（按天）
Posted
技术标签:
【中文标题】查询过去一年的活跃订阅人数和订阅人数（按天）【英文标题】：Query for Number of Active Subscribers and Subscriptions in Past Year (By Day) 【发布时间】：2015-06-23 15:36:39 【问题描述】：我正在寻求编写一个查询，以使我能够捕获过去一年（按天）的活跃订阅者和订阅数。例如，下面的查询将告诉我们在 2013 年 6 月 2 日至 2014 年 6 月 1 日期间任何时候主动订阅的唯一订阅者和订阅者的数量。
SELECT count(distinct(user_id)) as subscribers,
count(distinct(subscription_id)) as subscriptions
FROM dataset
WHERE (start_date <= to_date('01JUN2014:23:59:59','DDMONYYYY:HH24:MI:SS') and
end_date >= to_date('02JUN2013:00:00:00','DDMONYYYY:HH24:MI:SS')) and
payment_amt > 0;
话虽如此，我正在寻找一个查询，该查询允许我同时捕获过去 12 个月内活跃的所有活跃订阅者和活跃订阅（过去 365 天的按天）。我正在寻找的最终数据集看起来像这样（365 行；3 列）：
Ref_Date Subscribers Subscriptions
6/1/2014 100 150
6/2/2014 101 153
… … …
5/30/2015 106 160
5/31/2015 104 155
此外，对于本文中的查询，上表中“Ref_Date”的值将是“6/1/2014”。我只想用一个数据拉取所有 365 个“Ref_Date”。 任何人都可以提供任何帮助吗？非常感谢！
【问题讨论】：
【参考方案1】：您需要创建一个日期表并将其连接到数据中。下面的示例使用来自同一数据集的派生表来生成日期列表。
查询：
SELECT Ref_date
,count(DISTINCT (user_id)) AS subscribers
,count(DISTINCT (subscription_id)) AS subscriptions
FROM (
SELECT DISTINCT start_date AS ref_date
FROM dataset
WHERE start_date BETWEEN DATEADD(YYYY,-1,CURRENT_DATE) and CURRENT_DATE
) dateslist
INNER JOIN dataset ON start_date < DATEADD(YYYY, 1, ref_date)
AND end_date >= ref_date
【讨论】：
【参考方案2】：SQL Fiddle
Oracle 11g R2 架构设置：
CREATE TABLE dataset ( user_id, subscription_id, start_date, end_date, payment_amt ) AS
SELECT 1,1, SYSDATE - 370, SYSDATE - 370 + INTERVAL '1' MONTH, 10 FROM DUAL
UNION ALL SELECT 2,1, SYSDATE - 350, SYSDATE - 350 + INTERVAL '1' MONTH, 10 FROM DUAL
UNION ALL SELECT 3,2, SYSDATE - 345, SYSDATE - 345 + INTERVAL '1' MONTH, 10 FROM DUAL
UNION ALL SELECT 4,3, SYSDATE - 310, SYSDATE - 370 + INTERVAL '1' MONTH, 10 FROM DUAL
UNION ALL SELECT 5,4, SYSDATE - 250, SYSDATE - 250 + INTERVAL '1' MONTH, 10 FROM DUAL
UNION ALL SELECT 6,5, SYSDATE - 250, SYSDATE - 250 + INTERVAL '1' MONTH, 10 FROM DUAL
UNION ALL SELECT 7,6, SYSDATE - 240, SYSDATE - 240 + INTERVAL '1' MONTH, 10 FROM DUAL
UNION ALL SELECT 8,6, SYSDATE - 230, SYSDATE - 230 + INTERVAL '1' MONTH, 10 FROM DUAL
UNION ALL SELECT 9,7, SYSDATE - 100, SYSDATE - 100 + INTERVAL '1' MONTH, 10 FROM DUAL
UNION ALL SELECT 10,8, SYSDATE - 50, SYSDATE - 50 + INTERVAL '1' MONTH, 10 FROM DUAL
UNION ALL SELECT 11,9, SYSDATE - 20, SYSDATE - 20 + INTERVAL '1' MONTH, 10 FROM DUAL
UNION ALL SELECT 12,9, SYSDATE, SYSDATE + INTERVAL '1' MONTH, 10 FROM DUAL;
查询 1：
WITH all_dates_of_year AS (
SELECT TRUNC(SYSDATE) - INTERVAL '1' YEAR + LEVEL - 1 AS day
FROM DUAL
CONNECT BY
TRUNC(SYSDATE) - INTERVAL '1' YEAR + LEVEL - 1 <= TRUNC(SYSDATE)
)
SELECT day,
COUNT( DISTINCT user_id ) AS num_users,
COUNT( DISTINCT subscription_id ) AS num_subs
FROM all_dates_of_year a
LEFT OUTER JOIN
dataset d
ON ( a.day BETWEEN TRUNC( d.start_date ) AND TRUNC( d.end_date ) )
GROUP BY
a.day
Results：
| DAY | NUM_USERS | NUM_SUBS |
|-----------------------------|-----------|----------|
| June, 23 2014 00:00:00 | 1 | 1 |
| June, 24 2014 00:00:00 | 1 | 1 |
| June, 25 2014 00:00:00 | 1 | 1 |
| June, 26 2014 00:00:00 | 1 | 1 |
| June, 27 2014 00:00:00 | 1 | 1 |
| June, 28 2014 00:00:00 | 1 | 1 |
| June, 29 2014 00:00:00 | 1 | 1 |
| June, 30 2014 00:00:00 | 1 | 1 |
| July, 01 2014 00:00:00 | 1 | 1 |
| July, 02 2014 00:00:00 | 1 | 1 |
| July, 03 2014 00:00:00 | 1 | 1 |
| July, 04 2014 00:00:00 | 1 | 1 |
| July, 05 2014 00:00:00 | 1 | 1 |
| July, 06 2014 00:00:00 | 1 | 1 |
| July, 07 2014 00:00:00 | 1 | 1 |
| July, 08 2014 00:00:00 | 2 | 1 |
| July, 09 2014 00:00:00 | 2 | 1 |
| July, 10 2014 00:00:00 | 2 | 1 |
| July, 11 2014 00:00:00 | 2 | 1 |
| July, 12 2014 00:00:00 | 2 | 1 |
| July, 13 2014 00:00:00 | 3 | 2 |
| July, 14 2014 00:00:00 | 3 | 2 |
| July, 15 2014 00:00:00 | 3 | 2 |
| July, 16 2014 00:00:00 | 3 | 2 |
| July, 17 2014 00:00:00 | 3 | 2 |
| July, 18 2014 00:00:00 | 3 | 2 |
| July, 19 2014 00:00:00 | 2 | 2 |
| July, 20 2014 00:00:00 | 2 | 2 |
| July, 21 2014 00:00:00 | 2 | 2 |
| July, 22 2014 00:00:00 | 2 | 2 |
| July, 23 2014 00:00:00 | 2 | 2 |
| July, 24 2014 00:00:00 | 2 | 2 |
| July, 25 2014 00:00:00 | 2 | 2 |
| July, 26 2014 00:00:00 | 2 | 2 |
| July, 27 2014 00:00:00 | 2 | 2 |
| July, 28 2014 00:00:00 | 2 | 2 |
| July, 29 2014 00:00:00 | 2 | 2 |
| July, 30 2014 00:00:00 | 2 | 2 |
| July, 31 2014 00:00:00 | 2 | 2 |
| August, 01 2014 00:00:00 | 2 | 2 |
| August, 02 2014 00:00:00 | 2 | 2 |
| August, 03 2014 00:00:00 | 2 | 2 |
| August, 04 2014 00:00:00 | 2 | 2 |
| August, 05 2014 00:00:00 | 2 | 2 |
| August, 06 2014 00:00:00 | 2 | 2 |
| August, 07 2014 00:00:00 | 2 | 2 |
| August, 08 2014 00:00:00 | 2 | 2 |
| August, 09 2014 00:00:00 | 1 | 1 |
| August, 10 2014 00:00:00 | 1 | 1 |
| August, 11 2014 00:00:00 | 1 | 1 |
| August, 12 2014 00:00:00 | 1 | 1 |
| August, 13 2014 00:00:00 | 1 | 1 |
| August, 14 2014 00:00:00 | 0 | 0 |
| August, 15 2014 00:00:00 | 0 | 0 |
| August, 16 2014 00:00:00 | 0 | 0 |
| August, 17 2014 00:00:00 | 0 | 0 |
| August, 18 2014 00:00:00 | 0 | 0 |
| August, 19 2014 00:00:00 | 0 | 0 |
| August, 20 2014 00:00:00 | 0 | 0 |
...
| June, 21 2015 00:00:00 | 1 | 1 |
| June, 22 2015 00:00:00 | 1 | 1 |
| June, 23 2015 00:00:00 | 2 | 1 |
【讨论】：
以上是关于查询过去一年的活跃订阅人数和订阅人数（按天）的主要内容，如果未能解决你的问题，请参考以下文章