At 12:52 PM +0100 8/9/01, Dave Watkinson wrote:
>
>I have (many many many) tables ... two of which are linked by a third,
>so that there can be a one-to-many relationship. What I'd like to do is
>find which ids from table 1 have more than one relation in table 2, via
>table 3.
>
>I've tried this...
>
>SELECT empid, COUNT(empid) FROM emp_cont , employers, contacts WHERE
>emp_cont.empid = employers.uid and emp_cont.contid = contacts.uid
>and count(empid) > 1;
>
Your getting a lot of errors, because the statement has well a lot of
mistakes. Let's start near the beginning, you state:
SELECT empid, COUNT(empid) FROM emp_cont , employers, contacts
This statement should produce an error, since you didn't tell the
database which table to take empid from. A working statement would
look like
SELECT empid, COUNT(empid) FROM emp_cont or
SELECT emp_cont.empid, COUNT( emp_cont.empid) FROM emp_cont ,
employers, contacts
Either statement provides SQL with implicit instructions, as to which
empid to use and where.
The next major error is to use a Count function without the group by.
I'm not an SQL guru, but I believe whenever you use count you must
also use group by or having so adding a
SELECT empid, COUNT(empid) FROM emp_cont GROUP BY empid
will correct that.
Unfortunately, I think the larger issue is how your trying to find
your query. Your query did confuse me a little, so I could be wrong,
but technically speaking, if you are trying to locate an employee
with more than one record in employee contracts, then there's no need
to involve the contracts table within the query.
So say I have table Employees that contains empid, name, ssn. Then I
have table Employee Contracts that contains, empid, contractid.
Assuming that all empid(s) on Employee Contracts have a correlating
empid in Employees (which they will if the table was defined
properly) then the steps to get what you want are:
Step 1: Return a Count of empid
SELECT COUNT(empid) AS countid FROM emp_cont GROUP BY empid
Step 2: Return employees with more than one contract
SELECT empid FROM emp_cont WHERE countid >1
You'll note that both step requires a return, which general means a
subquery, I believe for MySQL to hanle subqueries you actually have
to perform to separate queries. If your not using MySQL then you
could write a statement such as:
SELECT empid FROM emp_cont WHERE (SELECT COUNT(empid) FROM emp_cont
GROUP BY empid) [note statement may need tweaking since I use MySQL
and can't test it]
If your using MySQL, I believe you need to return the results from
one, then use the results to find the next. Or you could use PHP to
cheat, and return all results, but only display the ones where the
count is greater than 1.
A subquery is best if your tables are large, returning 1000 results
to display 50 wastes time. But if you have under 200 employees, then
just use the SQL statement
$results=mysql_query (
SELECT empid, COUNT(empid) AS countid FROM emp_cont GROUP BY empid);
while ($row=mysql_fetch_array($results)){
$countid = $row["countid"];
$empid = $row["empid"];
if($countid=="1")
{
$display_employee .="";
}
else
{
$display_employee .="
<p>$empid</p>";
};
}
?>
<? echo "$display_employee"; ?>
Alnisa
--
.........................................
Alnisa Allgood
Executive Director
Nonprofit Tech
(ph) 415.337.7412 (fx) 415.337.7927
(url) http://www.nonprofit-techworld.org
(url) http://www.nonprofit-tech.org
(url) http://www.tech-library.org
.........................................
Nonprofit Tech E-Update
mailto:[EMAIL PROTECTED]
.........................................
applying technology to transform
.........................................
--
PHP Database Mailing List (http://www.php.net/)
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
To contact the list administrators, e-mail: [EMAIL PROTECTED]
