MySQL LEFT JOIN duplicates results

Question

I have a problem with implementing a module where one project can belong to multiple categories. Example: project \"PHP Programmer\" belongs to cat

Answer 1

You can also rewrite this as an "IN" to get around duplicates:

SELECT projects.*
FROM projects      
where projects.id in (select project_id
                      from projects_category
                      WHERE pojects_category.category_id IN (1,3,11)
                     ) and
      projects.id='94'

The "in" prevents duplicates from forming when you are using joins for filtering records.

Answer 2

No, this is fine. This is just one of the rare cases when you want to use the DISTINCT key word to remove duplicates.

In this case this is justified by the fact that the logic of the query is correct, even though it returns more than one row. Many times one can see the usage of DISTINCT when actually the logic of the query is wrong.

Side-note:

• any filter on a table reference you're using in the WHERE clause other than IS NULL/IS NOT NULL would make any LEFT JOIN on this same table reference turn to an INNER JOIN, as for the final resultset behaviour. (see this: https://stackoverflow.com/a/15483895/1291428)

• you ought not use GROUP BY to simulate the effect of DISTINCT, for 2 reasons:

  1/ This is just not the purpose. One of the effects of GROUP BY is to eliminate duplicates, but its main purpose is to group rows according to a certain set of criteria, in order to apply some analytic calculations/operations on them.

  2/ GROUP BY also ORDER BY the results (in mysql), which is not necessarly what you want and in that case slows down the execution. Please, just ensure appropriate use of what the engines are providing, that's always better from the point of view of forward compatibility. (anticipating that what you include as granted is actually not)

regards.