Question

The title might have been a bit confusing but I can clarify here.

I’m trying to use mysql’s LIKE to match a value that sometimes is surrounded by a : or sometimes it can be on one side, or sometimes neither. I know of the wildcard % but that will select everything, so I found the _ which would work but it ALWAYS select that single character, I want to only select that character if it is a : otherwise not.

Still confused? Here’s an example.

Say I got a column called ‘name’ in the table ‘que’ and in the column ‘name’ I got a value of ‘sdfgsdfgsdfg:jimmy:asfds’ I want to select ‘:jimmy:’ but if instead the value was ‘sdfgsdfgsdfg:jimmy’ I want to select ‘:jimmy’ only.

Another example.

To acomplish this with regular expressions it would be like this (:?)jimmy(:?) if that helps.

Whats the purpose?

I want to remove jimmy from that value and replace it with a :
I’m trying to store Multiple name into one column and need to remove that one person.

I got something like this so far:

 UPDATE `que` SET `name`= CASE WHEN `name` LIKE '%jimmy%' THEN replace(`name`, '%jimmy%', ':') ELSE `name` end WHERE 1

Answer 1

To acomplish this with regular expressions it would be like this (:?)jimmy(:?) if that helps.

So use REGEXP?

I’m trying to store Multiple name into one column and need to remove that one person.

Don’t do that. Use your RDBMS’s relational capabilities and normalise your schema:

CREATE TABLE que_names (
  que_id BIGINT UNSIGNED NOT NULL,
  name   VARCHAR(255)    NOT NULL,
  FOREIGN KEY (que_id) REFERENCES que (que_id)
);

INSERT INTO que_names
  (que_id, name)
VALUES
  (1, 'bob'),
  (1, 'fred'),
  (1, 'james'),
  (2, 'michael'),
  (2, 'jimmy'),
  (2, 'fred')
  -- etc.
;

ALTER TABLE que DROP COLUMN name;

You would then remove all jimmy with:

DELETE FROM que_names WHERE name = 'jimmy';

If so desired, you could then retrieve your existing table as follows:

SELECT   que.*, GROUP_CONCAT(que_names.name SEPARATOR ':') AS name
FROM     que LEFT JOIN que_names USING (que_id)
GROUP BY que_id