我正在尝试查找具有枚举列的所有行,该枚举列为NULL或与给定的枚举值集不同。我可以通过许多IS DINSTINCT FROM
调用来做到这一点,但是它确实很冗长,我宁愿使用NOT IN()
语法,但是NULL
会把它扔掉。
这里是一个我想在SQL中执行的操作的示例,它来自这个小提琴:http://sqlfiddle.com/#!17/dfae4d/8
CREATE TYPE mood AS ENUM ('sad', 'ok', 'happy');
CREATE TABLE people (
name text,
current_mood mood,
current_mood_as_text varchar
);
insert into people values('Mr Happy', 'happy', 'happy');
insert into people values('Mr Sad', 'sad', 'sad');
insert into people values('Mr Null', NULL, NULL);
-- This doesn't return MrUnknown because of the NULL value:
select * from people where current_mood NOT IN ('happy');
-- This works great, and includes Mr Null:
select * from people where current_mood IS DISTINCT FROM 'happy';
-- But If I want to start comparing to multiple moods, it gets more verbose fast:
SELECT * FROM people
WHERE current_mood IS DISTINCT FROM 'happy' AND
current_mood IS DISTINCT FROM 'sad';
-- You can't write this, but it's kinda what I want:
-- SELECT * FROM people
-- WHERE current_mood IS DISTINCT FROM ('happy', 'sad');
-- For the non enum column, I could do this to make my intention and code clear and concise
SELECT * FROM people
WHERE coalesce(current_mood_as_text, '') NOT IN ('happy', 'sad');
-- But if I write this, I get an error because '' is not a valid enum value
-- SELECT * FROM people
-- WHERE coalesce(current_mood, '') NOT IN ('happy', 'sad');
还有另一种方法可以使这种多重比较更加简洁吗?
答案 0 :(得分:1)
使用coalesce()
的解决方案:
select *
from people
where coalesce(current_mood not in ('happy', 'sad'), true)
答案 1 :(得分:0)
另一种方式:
SELECT * FROM people
WHERE NOT EXISTS (
SELECT FROM (VALUES ('happy'::mood), ('sad')) v(m)
WHERE people.current_mood = v.m
);
答案 2 :(得分:0)
最干净(IMO)的方法是在CTE中使用VALUES()
并对该表表达式执行NOT EXISTS()
:
WITH m(m) AS ( VALUES( 'happy'::mood) )
SELECT *
FROM people p
WHERE NOT EXISTS (
SELECT * FROM m
WHERE m.m = p.current_mood
);