Question

I want to do something like this

INSERT INTO t (t.a, t.b, t.c)  VALUES ('key1','key2','value')  ON DUPLICATE KEY UPDATE  t.c = 'value'; INSERT INTO t (t.a, t.b, t.c)  VALUES ('key1','key3','value2')  ON DUPLICATE KEY UPDATE  t.c = 'value2'; 

t.a and t.b are keys. This all works fine but i get an error on the second insert. With phpMyAdmin a query like this works fine but i’m guessing it’s running the queries independently as it prints out the results from that query as comments?

Something like this would be good too but i will need to have different values for each item. I prefer this but i’m not sure how i can change the value on the update for each value.

INSERT INTO t (t.a, t.b, t.c) VALUES ('key1','key2','value'), ('key1','key3','value2') ON DUPLICATE KEY UPDATE t.c = ??? 

The problem is in the question marks, what should i put there so that each insert/update will have the correct value? Obviously if i put a value there all the fields will get that value.

If there is another way of doing an ‘update if exists, otherwise insert’ query on multiple fields with two keys, i’m up for other ideas too. I guess i could run each query separately (like phpMyAdmin?) but it’s going to be a lot of queries so i really want to avoid that.

Answer 1

Use the VALUES() function

INSERT INTO t (t.a, t.b, t.c) VALUES ('key1','key2','value'), ('key1','key3','value2') ON DUPLICATE KEY UPDATE t.c = VALUES(t.c) 

see http://dev.mysql.com/doc/refman/5.0/en/insert-on-duplicate.html