Question

Suppose I have two tables, one with blog posts and another with readers and their comments:

Table 1:

table name: BlogPosts:

structure:

id (int)

title (string)

Table 2:

table name: Readers:

id (int)

blog_post_id (int)
name (string)
comment (string)

in the Readers table there is a unique composite key on blog_post_id/name (i.e. one comment per reader per post), though this may not matter for my question.

I want to be able to do a single query tells me what a particular reader’s comment was on each BlogPost, but it should include BlogPosts for which there is no comment entered for that reader (so the query should return one row for each blog post in the database).

I tried several variations that look like this:

SELECT  
     BlogPosts.id,
     BlogPosts.title,
     Readers.name,
     Readers.comment
    FROM
         BlogPosts
    RIGHT JOIN Readers ON
    (BlogPosts.id = Readers.blog_post_id)
    WHERE Readers.name = "joe"

..this just returns the rows where there is actually a comment from joe. Other variations where I was able to get all of the blog posts gave me an invalid identifier whenever I included the where clause.

I am using Oracle Express 10g in case that makes any difference.

Thanks for any help.

Answer 1

Everything looks correct except that you want to move the WHERE clause into the join, like this:

    SELECT BlogPosts.id,
           BlogPosts.title,
           Readers.name,
           Readers.comment
      FROM BlogPosts
RIGHT JOIN Readers ON BlogPosts.id = Readers.blog_post_id
       AND Readers.name = 'joe'

You effectively make the OUTER JOIN into an INNER one when you put the Readers.name = 'joe' in the WHERE clause.

Edit:
Thanks to the OP for clarification in the comments.
What you describe in your most recent comment can be achieved simply by switching from a RIGHT JOIN to a LEFT JOIN (as suggested earlier on by another commenter).
Note that you can get more than one row per BlogPost when there are multiple comments per post.