Question

Is it possible to write a function within mySQL that checks the value of a SELECT query, and if it != 0, continue running the function and add to the result array?

I have a few entries marked as related_id = 0 The rest have actual positive numbers attached.

I am trying to create a single query that will continually loop through a sql function, check and see if the retrieved related_id result is 0. If not, then it will run again, add the newly checked row to the result array and continue until it finds a 0.

The way the DB is set up, a 0 will always occur, and it will be searching via id,so there should only be one result.

I am using a slightly modified tagging system.

keywords_id | keyword  | related_id
------------------------------------
1           | sports   | 0
2           | baseball | 1
3           | football | 1
4           | pitchers | 2

If I pass in related_id = 2 I’m looking to return an array that yields:

array 
[0]
   keyword => 'sports'
   keywords_id => '1'
   related_id => '0'
[1]
   keyword => 'baseball'
   keywords_id => '2'
   related_id => '1'
[2]
   keyword => 'pitcher'
   keywords_id => '4'
   related_id => '2'

I tried my luck with using IF and THEN but i just haven’t seen enough examples and its tough to find solid tutorials online. Thanks!

SELECT k.* 
FROM keywords k
WHERE keywords_id = 2

The above query would essentially start it off. The keywords_id is the related_id that is passed. Therefore the above would return keywords_id: 2 |keyword: baseball |related_id: 1

Because related_id 1 is NOT 0, i would need to run it again.

I realize this can be done in PHP through a few extra queries…is that the way I should do this? I assume it could be done in mySQL after reading about functions. I just couldn’t figure it out.

Answer 1

MySQL doesn’t support returning an array from a function. Would suggest creating a table (or possibly a temporary table) to hold the results and then writing a stored procedure to populate it instead.

Not sure exactly what you are trying to achieve but maybe something along the lines of this?

CREATE TABLE results (keywords_id INT, keyword VARCHAR(2000), related_id INT);

CREATE PROCEDURE populateResults(IN keyword_id_in INT)
BEGIN
  DECLARE num_results INT;
  DECLARE max_results INT;
  DECLARE next_keyword_id INT;
  SET next_keyword_id = keyword_id_in;
  SET max_results = 5;
  SET num_results = 0;

  DELETE FROM results;

  the_loop: LOOP
    INSERT INTO results
    SELECT keywords_id, keyword, related_id
    FROM keywords
    WHERE keywords_id = next_keyword_id;

    SELECT related_id
    FROM keywords
    WHERE keywords_id = next_keyword_id
    INTO next_keyword_id;

    IF (next_keyword_id = 0) || (num_results >= max_results) THEN
      LEAVE the_loop;
    END IF;
    SET num_results = num_results + 1;
  END LOOP the_loop;
END

See SQL Fiddle Demo