Question

Many times just find the right article and reading it on StackOverflow helped me
through a lot of stuff that I didn’t know what to do with but for the first time,
I think I need to write one.
I’ve been searching for the correct answer on this prob but I’ve failed to find one.

So the bugger is that I’ve created 2 tables and one intersection table to relate them.

The idea behind is so simple(I’m almost embarrassed that I can’t solve this one).

One BBS article MAY have SOME attached files.
That means one article could have an attachment or not.
One article can have multiple attachments.

What I’ve been trying to do is

get the list of articles with all the information on its attachments but without duplicate rows.

Ok… I tried to put up the DDL but I coulndn’t format it correctly…

create table article(
id PK
some other stuff...
)

create table attachment(
id PK
physical_file_name 
etc...
)

and here is the intersection

create table article_attachment(
id PK(synthetic)
article_id FK
attachment_ID FK
)

I want to select all the articles whether it has or hasn’t any attachments
but if one article has multiple attachments I only need one of the attachments.
(it doesn’t matter which one of them)

Yes it sounds dumb but there is no DBA or SQL developer around here so I have to do all the
things… pretty messed up I’m trying my best

any wise ideas?

Thanks in advance

-p.s. –
I’ve tried something like…

with refined_table as(

  select file_id, row_number() over(partition by id order by id desc) as seq
  from consumer_file

)
select * 
from consumer_info ci 
left outer join consumer_file cf on cf.consumer_id = ci.id
left outer join refined_table rt on rt.file_id = cf.file_id
where rt.seq =1

this but I don’t really understand how it works

UPDATE

Here is what I tried at first. It keeps giving me
ORA-00979:not a GROUP BY expression
I looked for the solution for a long time. A few people suggested using
aggregation function or “Hint” (if that’s what it is…)

  with refined_table as(

      select file_id, row_number() over(partition by id order by id desc) as seq
      from consumer_file

    )

I mean this

Hmm any ideas ?

Many thanks all of you : )

SELECT 
    CI.id as article_id, 
    DF.id as attachment_id,  
    DF.PHYSICAL_NAME as file_name 
FROM 
    CONSUMER_INFO CI
LEFT OUTER JOIN 
    CONSUMER_FILE CF ON CF.CONSUMER_ID = CI.ID
LEFT OUTER JOIN 
    DEFAULT_FILE DF ON CF.FILE_ID = DF.id
GROUP BY 
   CI.ID

Answer 1

If its only 1->N relationship from article -> attachment then you can modify the attachment as follows:

create table attachment(
   id PK
   article_id FK 
   physical_file_name 
   etc...
)

You need a separate intersection table only if there were an N->N relationship between articles and attachments.

UPDATE: However, if you still need to keep the N<->N relationship (as mentioned in your comment), then you can use the GROUP BY clause to GROUP results by article_id e.g:

SELECT 
    article.id as article_id, 
    attachment.id as attachment_id,  
    attachment.physical_file_name as file_name 
FROM 
    article 
LEFT OUTER JOIN 
    article_attachment ON article_id = article.id 
LEFT OUTER JOIN 
    attachment ON attachment_id = attachment.id
GROUP BY 
    article_id