Re: Difference between inner join and join in SQL query

[Vern Hamberg <vhamberg@xxxxxxxxxxxxxxx>]

​​​​​Javier

LEFT JOIN is not an INNER JOIN and not the same as JOIN - according to 
the IBM i SQL Reference, there are 4 groups of JOIN operators - INNER, 
OUTER, EXCEPTION, and CROSS. The manual says - and Jerry, please do 
google searches or get the SQL Reference, all your answers are there.

The manual says that if there is no JOIN OPERATOR it is an INNER JOIN.

LEFT, RIGHT, or FULL by itself is an OUTER. You can't use OUTER by itself.

EXCEPTION by itself is a LEFT EXCEPTION, you can also specify RIGHT 
EXCEPTION.

Finally you have a CROSS.

These are very much SQL standard - IBM i is one of the most 
standards-compliant version of SQL in the industry.

Cheers
Vern

On 7/15/2022 1:01 PM, Javier Sanchez wrote:
> Okay, this is it.  JOIN and INNER JOIN are just the same.  Don't know why
> the SQL standards took that for granted.  LEFT JOIN also would work just
> fine as well as JOIN as long as you know what you are doing.  It's
> confusing, isn't it.  Okay, let's get back.  INNER JOIN is the same as
> JOIN.  LEFT JOIN would also be the same thing as JOIN.  But putting the
> OUTER JOIN construct would give you different results depending on what
> you want to do.
>
> Suppose you have two tables A, and B.  Now, if you select ALL records in
> table A that correspond to the same rows as in table B related with some
> key columns, that is called a cartesian relationship.
> One to one relationship would give you all records in table A that have the
> same key values as in table B.  JOIN or LEFT JOIN  or simply JOIN  would
> work well.
>
> Now, here's the difference.  Case 1.  We want back all rows in table A and
> also all those in table B that have the same key columns in the ON xxx =
> yyy part of the SQL statement, but want also to include those rows in table
> B that DO NOT match the key criteria.  Those rows will be returned to you
> with NULL values.  So you'll know they are not in table B.
>
> Case 2.  RIGHT OUTER JOIN works in the opposite way.  You will get all rows
> in table B as well as in table A that coincide with the key columns but
> you'll also get NULL rows in table A that ARE NOT in table B.  So you will
> know what is missing in table A that is not in table B.
>
> Got it?
>
> Javier.