Question

if the data in table is :

select * from data;

 key  | value                    
------+---------
 a    | a_value
 b    | b_value
 c    | c_value
 d    | d_value
 e    | e_value
 f    | f_value

if one client connects to my web server and ask for the value of a, my application will try to do:

select value where key = 'c';

and send the result value to client

if 3 clients connect to my web server, there are 3 statements occur,

  select value where key = 'e';  # client 1

  select value where key = 'c';  # client 2

  select value where key = 'a';  # client 3

for resource reason, I want to consolidate many statements into one select statement.

select * where key = 'e' or key = 'c' or key = 'a';

 key  | value                    
------+---------
 a    | a_value
 c    | c_value
 e    | e_value

but the problem is the sequence of result is not the same as my WHERE clause.
so I cannot distinguish the result from the clients’ query.
if I want to send value back to the clients

'e_value' to client 1
'c_value' to client 2
'a_value' to client 3

thanks!!

Answer 1

I don’t quite understand your use-case, but a query like the following enables you to

1. .. get results in pre-defined order
2. .. identify rows in result set

SELECT v.*, d.value
FROM   data
JOIN  (
   VALUES
     (1::int, 'e'::text)  -- explicit cast may not be necessary
    ,(2, 'c')
    ,(3, 'a')
    ) v(client, key) USING (key)
ORDER  BY v.client;

Returns:

client | key | value
-------+-----+------
1      | e   | e_value
2      | c   | c_value
3      | a   | a_value