Question

感觉就像一个愚蠢的问题。

我有以下查询

select 'Cardiac - ACS' as a
union 
select 'Cardiac - All' as a
order by a

我希望第二行“Cardiac - All”首先替换“连字符”字符。我该怎么替换它？（以及如何在Postgres中找出符号的顺序及其优先级？）

Answer 1

按字符排序

排序顺序取决于您的语言环境，尤其是collation。找到您的设置：

SHOW lc_collate;

在C语言环境中，字符根据其字节值进行排序。这些字符在连字符（'-'）之前排序：

SELECT chr(g)
FROM generate_series (33,ascii('-')) g

chr
---
!
"
#
$
%
&
'
(
)
*
+
,
-

对于其他排序规则，不为true。测试：

SELECT g, chr(g)
FROM   generate_series (33,ascii('-')) g
order  by chr(g) COLLATE "C"

连字符的十进制ASCII值为45.
移除COLLATE "C"以查看您的当前区域设置的效果。

要在排序规则中查找基本候选字符，请运行：

SELECT g, chr(g)
FROM   generate_series (1,255) g
ORDER  BY chr(g);

Answer 2

如果存在任意排序要求，则创建要排序的列：

select a, (left(a, position('-' in a)), arbitrary_order)
from (
    select 'Cardiac - ACS' as a, 9 as arbitrary_order
    union 
    select 'Cardiac - All', 0
) s
order by 2, 1
;
       a       |       row       
---------------+-----------------
 Cardiac - All | ("Cardiac -",0)
 Cardiac - ACS | ("Cardiac -",9)

arbitrary_order列的默认值类似于9或999，任意情况的值都会更低。

或测试是否存在' - All'字符串

select a, (
    left(a, position('-' in a)),
    substring(a from position('-' in a)) != '- All'
    )
from (
    select 'Cardiac - ACS' as a
    union 
    select 'Cardiac - All'
) s
order by 2, 1
;
       a       |       row       
---------------+-----------------
 Cardiac - All | ("Cardiac -",f)
 Cardiac - ACS | ("Cardiac -",t)

Postgresql - 通过正确的顺序

2 个答案:

按字符排序