Counting items in a view

Question

Hi all,
I have the following problem:
I have a table containing document data, each document has a doc_id as a unique indentifier. Each document is assigned to a project via the column proj_code:
CREATE TABLE DOCS
DOC_ID NUMBER(22) NULL,
PROJ_CODE VARCHAR2(10) NULL,
In another table I have the project codes, project names and customer names. Each project has one customer assigned to it, there can be multiple projects per customer. Project code and project name are in a 1:1 relationship:
CREATE TABLE PROJECT
PROJ_CODE NUMBER(22) NULL,
CUST_NAME VARCHAR2(30) NULL,
PROJ_NAME VARCHAR2(30) NULL
What I'm trying to do now is construct a view, that'll give me project code, project name, customer name and the count of documents, that have this project code. here's what I've come up with so far:
select proj_code, count(doc_id) from
(select project.proj_code, doc_id from project, docs where project.proj_code = docs.proj_code) group by proj_code;
This has two problems: it does not show the projects, that have no documents. Also, if I add the other columns to the select clause, I get this error:
ORA-00979: not a GROUP BY expression [SQL State=42000, DB Errorcode=979]
Ideas, anyone?
Thanks,
Mike

Answer 1

(+) is an Oracle specific way of signifying outer
joins. Since 9i Oracle has supported the ANSI
constructs LEFT OUTER and RIGHT OUTER, it's just that
old dogs such as myself tend to use (+) out of sheer
force of habit.I stick with the old syntax as to me left and right are totally confusing concepts when it comes to sets, and the new syntax is as verbose as anything you'd expect to be designed by a committee.
The big advanatge of the ANSI
version is that it also has FULL OUTER, which was
rather inconvenient to do in the earlier versions of
Oracle.Agreed. Though I've only ever written precisely one full outer join, and seen one other.