当对SQL UPDATE语句使用CASE时，为什么将NULL值输入数据库？

Question

To prevent multiple SQL statements from being sent to the database like this:

UPDATE `gmw-db-051514`.`gmw_provider`
SET `gmw_provider`.`is_partner` = TRUE,
`partner_url` = 'some-url',
`partner_remove_token` = 'ABC123'
WHERE `gmw_provider`.`provider_id` = '11';

UPDATE `gmw-db-051514`.`tbl_conversations`
SET `tbl_conversations`.`partner_initial_id` = 30
WHERE `tbl_conversations`.`user_initial_id` = 30;

UPDATE `gmw-db-051514`.`tbl_conversations`
SET `tbl_conversations`.`partner_target_id` = 30
WHERE `tbl_conversations`.`user_following_id` = 30;

... I would like to merge these SQL statements into one. I know that I can use IF and CASE in MYSQL statements but I think I'm using it wrong.

This is my attempt to use the CASE statement

UPDATE
  `gmw-db-051514`.`gmw_provider`,
  `gmw-db-051514`.`tbl_conversations`
SET
  `gmw_provider`.`is_partner` = true,
  `partner_url` = 'some-url',
  `partner_remove_token` = 'ABC123',
  `tbl_conversations`.`partner_initial_id` =
CASE WHEN 
  `tbl_conversations`.`user_initial_id` = '30' THEN '30' END,
  `tbl_conversations`.`partner_target_id` =
CASE WHEN 
  `tbl_conversations`.`user_following_id` = '30' THEN '30' END
WHERE
  `gmw_provider`.`provider_id` = '11'

The Condition works, but if there is already a value in a row, it is rewritten to null.

Have a look at this table:

 |user_initial_id|user_following_id|partner_initial_id|partner_target_id|
 |     30        |        41       |       NULL       |        41       |
 |     51        |        41       |       NULL       |        41       |
 |     37        |        30       |       NULL       |        NULL     |

After executing the UPDATE Statement it looks like this (41 is gone)

 |user_initial_id|user_following_id|partner_initial_id|partner_target_id|
 |     30        |        41       |       30         |        NULL     |
 |     51        |        41       |       NULL       |        NULL     |
 |     37        |        30       |       NULL       |        30       |

But it should look like this (41 still there and 30 added):

 |user_initial_id|user_following_id|partner_initial_id|partner_target_id|
 |     30        |        41       |       30         |        41       |
 |     51        |        41       |       NULL       |        41       |
 |     37        |        30       |       NULL       |        30       |

The row user_initial_id corresponds to partner_initial_id .

The row user_following_id corresponds to partner_target_id .

Like this:

        A                 B                 A                  B
 |user_initial_id|user_following_id|partner_initial_id|partner_target_id|

Can you help me out with this or another attempt? The "IF" statement leads to the same result.

Of course I used google and found similar questions here on stackoverflow, but the examples were more simple and not in combination with another table. I was afraid that this could affect the exact solution for these conditions.

If you can't help me, I use the multi-query version but this looks very ugly to me.

Thanks!

EDIT: My problem was solved and I've learned something new! If someone looks for a similar solution consider: 1.) Using Joins 2.) Don't be afraid of "sending a few queries through the same connection" :-)

Thanks to @dognose and @gordon Linoff!

Answer 1

You need an else clause for the case :

  `tbl_conversations`.`partner_initial_id` =
(CASE WHEN `tbl_conversations`.`user_initial_id` = '30' THEN '30' 
      ELSE `tbl_conversations`.`partner_initial_id` END)

I would also recommend table aliases and explicit JOIN conditions as well.