Question

Question:

What is the correct syntax to use when selecting a column value into an UPDATE/INSERT trigger NEW variable?

Details:

Here are the [simplified] tables I am using as well as the query. For the purposes of this example I have chosen something that is recursive, in this case a comment (ie. one can comment on a picture, but also on the comment). At any given time I want to know what picture a comment is commenting on, but I don’t want to query the table n times to get to that picture (hence why I am rejecting the Adjacency List Model), so I have a trigger on insert to copy the target picture over.

PICTURES table:

CREATE TABLE IF NOT EXISTS pictures
(
   id INT NOT NULL AUTO_INCREMENT,
   PRIMARY KEY ( id )
)

COMMENTS table:

CREATE TABLE IF NOT EXISTS comments
(
   id INT NOT NULL AUTO_INCREMENT,
   picture_id INT NOT NULL,
   comment_id INT NOT NULL,
   PRIMARY KEY ( id ),
   FOREIGN KEY ( picture_id ) REFERENCES pictures ( id ),
   FOREIGN KEY ( comment_id ) REFERENCES comments ( id )
)

COMMENTS trigger: Note that NEW.comment_id and NEW.picture_id are NOT NULL, therefore, they default to zero (FALSE). In my actual program I have an XOR to reject invalid attempts to assign both a picture and a comment id.

CREATE TRIGGER bi_comments_fer BEFORE INSERT ON comments FOR EACH ROW
BEGIN
   IF ( NEW.comment_id )
   THEN
      select NEW.picture_id := picture_id from comments where id = NEW.comment_id;
   END IF;
END

I tried including an @ sign, like so

select @NEW.picture_id := picture_id from comments where id = NEW.comment_id;

But that also raised a syntax error. Can anyone tell me how to fix this?

Answer 1

SET NEW.picture_id := (SELECT picture_id FROM comments WHERE id = NEW.commend_id);