Sql 具有所需顺序的联合查询

我正在使用Postgres，假设我有一个这样的简单表

cities
------------------
id | name
------------------
 1 | San Diego
------------------
 2 | San Antonio
------------------
 3 | Stockholm
------------------
 4 | Helsinki
------------------
 5 | San Francisco
------------------

 -------------------
 id | name
 -------------------
  5 | San Francisco
 -------------------
  1 | San Diego
 -------------------
  2 | San Antonio
 -------------------

 SELECT * FROM cities WHERE name = "San Francisco"
 UNION
 SELECT * FROM cities WHERE name ~* "San|Francisco"

我想找到所有完全或部分符合该模式的城市。重要的是，精确匹配的城市应该是结果的第一位。例如，如果我为“旧金山”找到城市，结果应该是这样的

cities
------------------
id | name
------------------
 1 | San Diego
------------------
 2 | San Antonio
------------------
 3 | Stockholm
------------------
 4 | Helsinki
------------------
 5 | San Francisco
------------------

 -------------------
 id | name
 -------------------
  5 | San Francisco
 -------------------
  1 | San Diego
 -------------------
  2 | San Antonio
 -------------------

 SELECT * FROM cities WHERE name = "San Francisco"
 UNION
 SELECT * FROM cities WHERE name ~* "San|Francisco"

重要的是，“旧金山”是结果中的第一个

我的问题是这样的

cities
------------------
id | name
------------------
 1 | San Diego
------------------
 2 | San Antonio
------------------
 3 | Stockholm
------------------
 4 | Helsinki
------------------
 5 | San Francisco
------------------

 -------------------
 id | name
 -------------------
  5 | San Francisco
 -------------------
  1 | San Diego
 -------------------
  2 | San Antonio
 -------------------

 SELECT * FROM cities WHERE name = "San Francisco"
 UNION
 SELECT * FROM cities WHERE name ~* "San|Francisco"

但是这个查询不能保证所需的顺序。我尝试使用分区添加行号（）并按其排序

WITH cities_with_rows_number AS (
   SELECT *, 0 as row_number FROM cities WHERE name = "San Francisco"
   UNION
   SELECT *, row_number() OVER() as row_number FROM cities WHERE name ~* "San|Francisco"
) SELECT * FROM cities_with_rows_number ORDER BY row_number

---

但在这种情况下，我不能轻易地删除重复的城市和“旧金山”将打印两次。我相信可以编写非常简单的查询来解决此任务。

您可以使用

order by

：

SELECT *
FROM cities
WHERE name ~* 'San|Francisco'
ORDER BY (name = 'San Francisco')::int DESC;

---

UNION

/

此查询不需要UNION ALL

。

您可以使用

order by

：

SELECT *
FROM cities
WHERE name ~* 'San|Francisco'
ORDER BY (name = 'San Francisco')::int DESC;

此查询不需要使用

UNION

/

UNION ALL