更新表連接的行

Question

OK，我有這樣的數據表（MyISAM數據）：

This is `table` 
    id | field1 | field2 | field3 | field4 | field5 
    1 | val1 | val2 | val3 | val4 | 1 
    2 | val1 | val2 | val3 | val4 | 1 
    3 | val2 | val1 | val4 | val3 | 1 
    4 | val2 | val1 | val4 | val3 | 1 

我需要更新這個表是這樣的（見我的WHERE條款）：

CREATE TEMPORARY TABLE IF NOT EXISTS `table_temp` 
AS 
(
    SELECT 
     * 
    FROM 
     `table` 
    WHERE 
     `field7` IS NULL 
     AND 
     `field8` IS NULL 
     AND 
     `field9` IS NULL 
); 

UPDATE LOW_PRIORITY IGNORE `table` AS `t`, `table_temp` AS `t2` 
SET 
    `t`.`field7`=CURRENT_TIMESTAMP(), 
    `t`.`field9`=`t2`.`id` 
WHERE 
    `t`.`field1` = `t2`.`field2` 
    AND 
    `t`.`field2` = `t2`.`field1` 
    AND 
    `t`.`field3` = `t2`.`field4` 
    AND 
    `t`.`field4` = `t2`.`field3` 
    AND 
    `t`.`field5` = `t2`.`field5` 
    AND 
    `t`.`field7` IS NULL 
    AND 
    `t`.`field8` IS NULL 
    AND 
    `t`.`field9` IS NULL 

查詢完成後

，我有這樣的：

This is `table` after query 
     id | field1 | field2 | field3 | field4 | field5 | field9 
     1 | val1 | val2 | val3 | val4 | 1  | **3** 
     2 | val1 | val2 | val3 | val4 | 1  | 
     3 | val2 | val1 | val4 | val3 | 1  | **1** 
     4 | val2 | val1 | val4 | val3 | 1  | 

但我需要這個（需要更新所有的4行）：

 id | field1 | field2 | field3 | field4 | field5 | field9 
     1 | val1 | val2 | val3 | val4 | 1  | 3 
     2 | val1 | val2 | val3 | val4 | 1  | **4** 
     3 | val2 | val1 | val4 | val3 | 1  | 1 
     4 | val2 | val1 | val4 | val3 | 1  | **2** 

實際上，所有的4行受到了影響，但使用IGNORE不只是一個簡單的關鍵字，field9有UNIQUE指數。

至於我能理解，我的查詢每次從同一行開始在t2 - >所以我有IGNORE和真的影響只有2行，而不是4.我的意思是，當系統即將更新第二排它從一開始看，第一個匹配的where子句是第3行 - >IGNORE，因爲當系統更新第1行時，第一個匹配行是id=3。我需要推動下一行的東西。

我嘗試設置另一個SET：

... 
SET 
    `t`.`field7` = CURRENT_TIMESTAMP(), 
    `t`.`field9` = `t2`.`id`, 
    `t2`.`field7` = CURRENT_TIMESTAMP(), 
    `t2`.`field9` = `t`.`id` 
... 

但它並沒有幫助我。

有沒有辦法完成我所需要的？

Answer 1

您需要更新兩次 - 1升，然後 2.降

CREATE TABLE Table1 
    (
     `id` int, 
     `field1` varchar(4), 
     `field2` varchar(4), 
     `field3` varchar(4), 
     `field4` varchar(4), 
     `field5` int, 
     `field7` int, 
     `field8` int, 
     `field9` int UNIQUE 
    ) 
; 

INSERT INTO Table1 
    (`id`, `field1`, `field2`, `field3`, `field4`, `field5`) 
VALUES 
    (1, 'val1', 'val2', 'val3', 'val4', 1), 
    (2, 'val1', 'val2', 'val3', 'val4', 1), 
    (3, 'val2', 'val1', 'val4', 'val3', 1), 
    (4, 'val2', 'val1', 'val4', 'val3', 1) 
; 

CREATE TEMPORARY TABLE IF NOT EXISTS `table_temp` 
AS 
(
    SELECT 
     * 
    FROM 
     `table1` 
    WHERE 
     `field7` IS NULL 
     AND 
     `field8` IS NULL 
     AND 
     `field9` IS NULL 
    order by id 
); 

UPDATE LOW_PRIORITY IGNORE `table1` AS `t`, `table_temp` AS `t2` 
SET 
    `t`.`field7`=CURRENT_TIMESTAMP(), 
    `t`.`field9`=`t2`.`id` 
WHERE 
    `t`.`field1` = `t2`.`field2` 
    AND 
    `t`.`field2` = `t2`.`field1` 
    AND 
    `t`.`field3` = `t2`.`field4` 
    AND 
    `t`.`field4` = `t2`.`field3` 
    AND 
    `t`.`field5` = `t2`.`field5` 
    AND 
    `t`.`field7` IS NULL 
    AND 
    `t`.`field8` IS NULL 
    AND 
    `t`.`field9` IS NULL 
; 

CREATE TEMPORARY TABLE IF NOT EXISTS `table_temp2` 
AS 
(
    SELECT 
     * 
    FROM 
     `table1` 
    WHERE 
     `field7` IS NULL 
     AND 
     `field8` IS NULL 
     AND 
     `field9` IS NULL 
    order by id desc 
); 

UPDATE LOW_PRIORITY IGNORE `table1` AS `t`, `table_temp2` AS `t2` 
SET 
    `t`.`field7`=CURRENT_TIMESTAMP(), 
    `t`.`field9`=`t2`.`id` 
WHERE 
    `t`.`field1` = `t2`.`field2` 
    AND 
    `t`.`field2` = `t2`.`field1` 
    AND 
    `t`.`field3` = `t2`.`field4` 
    AND 
    `t`.`field4` = `t2`.`field3` 
    AND 
    `t`.`field5` = `t2`.`field5` 
    AND 
    `t`.`field7` IS NULL 
    AND 
    `t`.`field8` IS NULL 
    AND 
    `t`.`field9` IS NULL 
; 

select * from table1 
;