Question

I have a multiple JOIN in the form of

SELECT * FROM column1
  LEFT JOIN column2 USING(id)
  JOIN column3 USING(name)
  JOIN column4 USING(info)
  WHERE column1.id = 44

If changing the LEFT JOIN to JOIN; for obvious reason, the query will return nothing in the absence of column2.id. The main data is column1 and the others are supportive data; thus, I want to retrieve data just if column1.id exists.

My question is: Does the first LEFT JOIN guarantees the query to return column1 values where column1.id exists, regardless of subsequent JOINs (e.g. when column3.name does not exist)? Or I need to change the next (all) JOINs to LEFT JOIN too?

Answer 1

You have a few options as the direct answer to your question is no.

Change all subsequent joins to LEFT JOINs.

SELECT * FROM column1
  LEFT JOIN column2 USING(id)
  LEFT JOIN column3 USING(name)
  LEFT JOIN column4 USING(info)
  WHERE column1.id = 44

If possible, re-arrange the query…

SELECT * FROM column1
  JOIN column3 USING(name)
  JOIN column4 USING(info)
  LEFT JOIN column2 USING(id)
  WHERE column1.id = 44

Or…

SELECT * FROM column1
  JOIN
  (
       column3
  JOIN column4 USING(info)
  LEFT JOIN column2 USING(id)
  )
    USING(name)
  WHERE column1.id = 44

On a side note, I’d avoid natural joins and the USING syntax. The logic is less transparent, it’s more open to bugs, more open to break later (if you add a field to one of the tables duplicating a field in the USING option), etc, etc.