将具有不同数据类型和列数的两个 SELECT 语句合并为 PostgreSQL 中的一个输出

Question

我有两个问题。第一个-

SELECT
  communications.creation_date as message_date, 
  message as message_text, 
  employees.first_name || ' ' || coalesce(employees.middle_name,'') || ' ' || employees.last_name as message_by
FROM app.communications 
INNER JOIN app.employees ON communications.message_from = employees.emp_id 
WHERE send_as_sms = TRUE AND com_id = (SELECT MAX(com_id) FROM app.communications)

基本上输出 - | message_date | message_text | message_by |

第二个查询 -

SELECT
  cs.com_id, 
  cs.first_name ||' ' || cs.last_name AS recipient_name, 
  cs.sim_number AS phone_number
FROM app.communication_sms cs
WHERE cs.com_id = (SELECT MAX(cs2.com_id) FROM app.communication_sms cs2)
ORDER BY first_name ASC

输出 - | com_id | recipient_name | phone_number |

从查询中可以看出，两个表都有一个 "com_id" 列。我需要的是进行一个查询，该查询将合并上面的两个查询以获得单个输出，例如 -

我怎样才能做到这一点？由于不同的数据类型和不同的列数，我不能使用 UNION。我会感谢你们的帮助。

Answer 1

您需要联合或加入的原因是什么？在上下文中，你问的问题 com_id 应该无关紧要。

尝试这样的事情，（这个查询没什么特别的，基本上只是将两者合并在一起）

SELECT
  communications.creation_date as message_date, 
  message as message_text, 
  employees.first_name || ' ' || coalesce(employees.middle_name,'') || ' ' || employees.last_name as message_by
  cs.com_id, 
  cs.first_name ||' ' || cs.last_name AS recipient_name, 
  cs.sim_number AS phone_number
FROM app.communications, app.communication_sms cs
INNER JOIN app.employees ON communications.message_from = employees.emp_id 
WHERE send_as_sms = TRUE AND com_id = (SELECT MAX(com_id) FROM app.communications)
AND cs.com_id = (SELECT MAX(cs2.com_id) FROM app.communication_sms cs2)
ORDER BY first_name ASC

Answer 2

不确定 com_id 是否相等，但如果它们可能不相等，我建议这样做：

select * -- list out the columns, I haven't bothered here
FROM (
      SELECT MAX(com_id) as com_id FROM app.communications 
      UNION
      SELECT MAX(cs2.com_id) FROM app.communication_sms cs2
      ) u
left join (
      SELECT
        com_id -- don't know which table this comes from
        communications.creation_date as message_date, 
        message as message_text, 
        employees.first_name || ' ' || coalesce(employees.middle_name,'') || ' ' || employees.last_name as message_by
      FROM app.communications 
      INNER JOIN app.employees ON communications.message_from = employees.emp_id 
      WHERE send_as_sms = TRUE AND com_id = (SELECT MAX(com_id) FROM app.communications)
     ) s1 on u.com_id = s1.com_id
left join (
SELECT
        cs.com_id, 
        cs.first_name ||' ' || cs.last_name AS recipient_name, 
        cs.sim_number AS phone_number
      FROM app.communication_sms cs
      WHERE cs.com_id = (SELECT MAX(cs2.com_id) FROM app.communication_sms cs2)
      ORDER BY first_name ASC
      ) s2 on u.com_id = s2.com_id

请注意，使用 CTE 可以避免少量重复

Answer 3

第一个查询： 您需要 MAX(c.com_id) 标识的最后一次通信，但前提是它是 SMS (c.send_as_sms = TRUE)。对于此 SMS，您需要发件人的姓名。此查询会生成一行或零行，具体取决于上次通信是否为 SMS。

第二个查询：您想要本次由MAX(cs.com_id)标识的最后一条短信通信的最后一条短信。这看起来多余。当你从存在匹配的 communication_sms 记录中看到这是一条短信时，为什么你在 communications 中有一个 send_as_sms？您可能需要考虑一下这个设计。无论如何，从这里你想得到收件人的名字加上一些 phone 号码。

如何结合这两个查询并不明显。您真的想要包含收件人姓名的最新 SMS，但只添加发件人姓名以防最后一条 SMS 也是最后一次通信吗？这似乎不太可能。我猜你更想

最新通讯，如果是短信或
最新短信通讯

对于发件人和收件人。或者您可以通过附加 SMS 进行非 SMS 通信吗？

这里查询最后一次短信通讯：

SELECT
  c.creation_date AS message_date, 
  c.message AS message_text, 
  e.first_name || ' ' || coalesce(e.middle_name,'') || ' ' || e.last_name AS message_by,
  cs.com_id, 
  cs.first_name ||' ' || cs.last_name AS recipient_name, 
  cs.sim_number AS phone_number
FROM (SELECT * FROM app.communication_sms ORDER BY com_id DESC FETCH FIRST ROW ONLY) cs
JOIN app.communications c ON c.com_id = cs.com_id
JOIN app.employees e ON c.message_from = e.emp_id 
ORDER cs.first_name ASC;

将具有不同数据类型和列数的两个 SELECT 语句合并为 PostgreSQL 中的一个输出

Merge two SELECT statements with different data types and number of columns into one output in PostgreSQL

sql

postgresql-9.1

multi-query