Postgres中的批處理條件INSERT INTO語句

Question

我想在Postgres中進行UPSERT ，但是我的版本低於9.5，所以我不能使用ON CONFILICT 。 我發現了這樣的事情：

UPDATE table SET field='C', field2='Z' WHERE id=3;
INSERT INTO table (id, field, field2)
       SELECT 3, 'C', 'Z'
       WHERE NOT EXISTS (SELECT 1 FROM table WHERE id=3);

但是我該如何使用批處理值呢？ 該語句可以表達我想做的事情（它不起作用）：

INSERT INTO result_data (HOST_ID, RESOURCE_TYPE, TIMESTAMP, UPPER, LOWER, AVG) 
    (values ("xxxxx", "cpu", 1544313900, 80, 70, 75), ("aaaaaa", "cpu", 1544314000, 80, 70, 75))
    AS result2(HOST_ID, RESOURCE_TYPE, TIMESTAMP, UPPER, LOWER, AVG) 
    WHERE NOT EXISTS(SELECT 1 FROM result_data WHERE result2.HOST_ID = result_data.HOST_ID);

Answer 1

在WITH子句（CTE）中使用多個VALUES ，並在INSERT INTO.. SELECT的SELECT部分中引用它INSERT INTO.. SELECT

with result2(HOST_ID, RESOURCE_TYPE, TIMESTAMP, UPPER, LOWER, AVG)  AS
(
 VALUES  ('xxxxx', 'cpu', 1544313900,  80, 70, 75 ), 
         ('aaaaaa', 'cpu', 1544314000, 80, 70, 75 ) 
)
 SELECT * FROM   result2 r2 
   WHERE  NOT EXISTS 
           ( 
             SELECT 1 
               FROM   result_data r 
              WHERE   r.host_id = r2.host_id )

DEMO