有沒有辦法返回按「WHERE IN」值排序的SELECT結果？

Question

說我有一個電話

SELECT * FROM tablename WHERE id IN (5,3,8,9); 

返回在3,5,8,9的ID順序的結果。有沒有辦法讓它返回給定的5,3,8,9順序的結果？

Answer 1

SELECT * FROM tablename WHERE id IN (5,3,8,9) 
ORDER BY 
CASE id 
    WHEN 5 THEN 1 
    WHEN 3 THEN 2 
    WHEN 8 THEN 3 
    WHEN 9 THEN 4 
    ELSE 5 
    END 

Fiddle

Answer 2

您還可以通過使用顯式join做到這一點：

with vals as (
     select 5 as id, 1 as ordering union all 
     select 3, 2 union all 
     select 8, 3 union all 
     select 9, 4 
    ) 
select t.* 
from tablename t join 
    vals v 
    on t.id = v.id 
order by v.ordering; 

這允許你一次創建列表。

Answer 3

另一個更靈活的選項，因爲你只需要更改一次ids。

with id_list (id, sort_order) as (
    values (5,1), (3,2), (8,3), (9,4) 
) 
select t.* 
from tablename 
    join id_list l on l.id = t.id 
order by l.sort_order; 

維護SORT_ORDER然而，如果你需要插入新的ID有點醜陋。

隨着即將到來的9.4，這將是更容易：

with id_list (id, sort_order) as (
    select * 
    from unnest(array[5,3,8,9]) with ordinality 
) 
select * 
from tablename 
    join id_list l on l.id = t.id 
order by l.sort_order; 

這可以使用前9.4來完成，但是這依賴這不保證給予UNNEST函數的順序一致的排序（但實際上它似乎總是相同的）：

with id_list (id, sort_order) as (
    select *, row_number() over() 
    from unnest(array[5,3,8,9]) 
) 
select * 
from tablename 
    join id_list l on l.id = t.id 
order by l.sort_order; 

Answer 4

在postgres中，使用POSITION：

SELECT * 
FROM tablename 
WHERE id IN (5,3,8,9) 
ORDER BY POSITION(',' || id || ',' IN ',5,3,8,9,') 

見SQLFiddle

---

在MySQL中，使用FIELD()功能：

SELECT * 
FROM tablename 
WHERE id IN (5,3,8,9) 
ORDER BY FIELD(id,5,3,8,9) 

見SQLFiddle

Answer 5

你試試這種替代也

select * 
from Table1 
where id in (5,3,8,9) 
order by position(id::text in '5,3,8,9') 

sqlfiddle

---

position

Answer 6

-- some data 
DROP SCHEMA tmp CASCADE; 
CREATE SCHEMA tmp ; 
SET search_path=tmp; 

CREATE TABLE tablename 
     (id serial PRIMARY KEY 
     , blah integer 
     ); 
INSERT INTO tablename(blah) 
select gs 
FROM generate_series(1,10) gs; 

-- The query 
SELECT * FROM tablename t 
JOIN (SELECT id 
     , row_number() OVER() AS rnk 
     FROM unnest(array[ 5,3,8,9] ) AS id 
     ) val USING(id) 
ORDER BY val.rnk 
     ; 

結果：

DROP SCHEMA 
CREATE SCHEMA 
SET 
CREATE TABLE 
INSERT 0 10 
id | blub | rnk 
----+------+----- 
    5 | 5 | 1 
    3 | 3 | 2 
    8 | 8 | 3 
    9 | 9 | 4 
(4 rows)