This question might be trivial or even silly but I was wondering if there is a way to use "IN" on more than one column on one to one matching.
For example I use
select emp_id from employee where emp_id IN (select emp_id from employee_other)
How could I achieve something like
select emp_id from employee where emp_id,emp_org IN (select emp_id,emp_org from employee_other)
I know I cant be using the following because it will simply do the union whereas I want a selection based on one to one record matching.
select emp_id from employee where emp_id IN (select emp_id from employee_other) and emp_org in (select emp_org from employee)
Please note that I am reluctant to use EXCEPT.
Thanks guys
You may want to use the EXISTS operator
select e.emp_id
from employee e
where EXISTS
(
SELECT *
FROM employee_other eo
WHERE e.emp_id = eo.emp_id
AND e.emp_org = eo.emp_org
)
IN in Microsoft SQL Server only works with a single column, ie. you can only write X IN (...), never anything remotely like X,Y IN (...).
There are two ways to handle this, depending on your data:
Joining with a sub-query
Using EXISTS
To JOIN, do this:
select emp_id
from employee
inner join (select emp_id,emp_org from employee) as x
on employee.emp_id = x.emp_id and employee.emp_org = x.emp_org
Your example is a bit lousy, however, since you're using the same table.
To use EXISTS, do this:
select emp_id
from employee
where exists (
select emp_id,emp_org from employee e2
where e2.emp_id = employee.emp_id and e2.emp_org = employee.emp_org)
This, in the same way as the join, links the main table to the "sub-query" table, but whereas the join will produce duplicate rows if the "sub-query" produces multiple hits, the EXISTS clause will not.
I don't understand what you are trying to accomplish with emp_org in (select emp_org from employee) isn't that always true?
does this work?
select emp_id from employee e
where exists (select 1 from employee_other eo
WHERE e.emp_id =eo.emp_id and
AND e.emp_org = eo.emp_org )
You had it almost completely right in your second example. You just need to add parens around your column names.
select emp_id
from employee
where (emp_id,emp_org) IN (select emp_id,emp_org from employee)
Use Inner Join
select e1.emp_id from employee e1
inner join employee_other e2 on e1.emp_id = e1.emp_id and e1.emp_org = e2.emp_org
You may have to use Distinct in case the employee_other table causes dups.
select Distinct e1.emp_id from employee e1
inner join employee_other e2 on e1.emp_id = e1.emp_id and e1.emp_org = e2.emp_org
Related
How do we use an aggregate function, like COUNT, with * in conjunction with an alias? COUNT allows table alias to be used with column names but not *.
This works:
select e.*, COUNT(e.MGR)
from EMP as e
left outer join EMP as e1 on e.EMPNO = e1.EMPNO
group by e.[EMPNO], e.[ENAME], e.[JOB], e.[MGR], e.[HIREDATE], e.[SAL], e.[COMM], e.[DEPTNO]
but this does not:
select e.*, COUNT(e.*)
from EMP as e
left outer join EMP as e1 on e.EMPNO = e1.EMPNO
group by e.[EMPNO], e.[ENAME], e.[JOB], e.[MGR], e.[HIREDATE], e.[SAL], e.[COMM], e.[DEPTNO]
This example being used above is a made up one to demonstrate the case. There is no NOT-NULL column in either table. So how do we count the number of rows in left hand side table of a left outer join
This answers the original version of the question.
Use either the primary key or one of the columns used in the JOIN:
select e.*, COUNT(e1.EMPNO)
There is no NOT-NULL column in either table. So how do we count the
number of rows in left hand side table of a left outer join
This should not happen, because every table should have a primary key, guaranteeing at least one NOT NULL column.
Without such a column you can create one at runtime by joining onto a derived table that projects a NOT NULL constant and counting that column.
select e.*, COUNT(e.Flag)
from (SELECT, *, 1 AS Flag FROM EMP) as e
left outer join EMP as e1 on e.EMPNO = e1.EMPNO
group by ...
I have two tables:
Employees (columns: ID, Name)
and
employee partners (EmployeeID1, EmployeeID2, Time)
I want to output EmployeName1, EmployeeName2, Time instead of imployee ids.
(In other words, replace the ids with names, but in two columns at a time)
How would I do this? Would JOIN be the appropriate command?
you need to join the employee table 2 times as the employee partners table acts as many to many connection.
The select should be:
SELECT emp1.name, emp2.name, em.time
FROM Employees emp1
JOIN employee_partners em ON emp1.id = EmployeeID1
JOIN Employees emp2 on emp2.id = EmployeeID2
Often in these situations, you want to use LEFT JOIN:
SELECT e1.name as name1, e2.name as name2, em.time
FROM employee_partners ep LEFT JOIN
Employees e1
ON e1.id = ep.EmployeeID1 LEFT JOIN
Employees e2
ON e2.id = ep.EmployeeID2;
Notes:
The LEFT JOINs ensure that you do not lose rows if either of the employee columns is NULL.
Use tables aliases; they make the query easier to write and to read.
Qualify all columns names; that is, include the table name so you know where the column is coming from.
I also added column aliases so you can distinguish between the names.
Suppose we have two tables
student(studentID, name, department_ID)
department(departmentID, name).
Our aim is to retrieve the data from both tables using the subquery. I'm trying this
select * from department, student
where department.departmentID
IN (select student.departmentID from student, department
where student.departmentID = department.departmentID)
but it returns the cross product of the number of rows of two tables.
This is possible to get the correct result using JOIN like this
select * from department
Inner join student
on student.departmentID = department.departmentID
and using WHERE clause like this
select * from department, student
where department.departmentID = student.departmentID
I'm wondering if someone can tell me how it can be possible using subquery in SQL.
Hope this helps:
select *, (select name from department d where s.departmentID = d.departmentID) as dname
from student s
where (select name from department d where s.departmentID = d.departmentID) is not null
This problem is however meant to be solved using joins. To learn subQ, use proper examples.
SQL Fiddle: Test
This may seems the most dumbest question ever on stackoverflow but I am just wondering why would you write such a query:
Select e1.Emploee_ID, e1.Departement_ID From Employee e
Inner join Employee E1 on e.employee_id= e1.Employee_ID
Inner join Departement d on d.dep_id= e1.departement_id
Why do we have to join on employee? my obvious query would be
select e.employee_ID, e.Departement_id from employee e
inner join Departement d on d.dep_id= e1.departement_id
Referencing the PK with an inner join is redundant.
You would normally join on the same table to link with another record, for example if you have a FK-column that reference the boss of an employee.
Assuming you would have a nullable foreign-key column Boss_ID in table Employee
Select e.Employee_ID, boss.Employee_id, d.Departement_ID
From Employee e
LEFT OUTER JOIN Employee boss on boss.Employee_ID=e.Boss_ID
INNER JOIN Departement d on d.dep_id= e.departement_id
Note that i've used a LEFT OUTER JOIN to get also the employee that have no bosses.
What I can see. You do not need this join.
Inner join Employee E1 on e.employee_id= e1.Employee_ID
The two queries will give the same result. I can not see the point of JOINing twice on the Employee table.
I can't see any reason to join Employee to Employee in this query. In the past I've occasionally used two subsets of the same table in the same query, but there's nothing like that going on here. To me it looks like this was done by mistake.
If employee_id is a PK then it doesn't make sense, but if it is not the two queries will return different results.
The first query will not return NULL employee_id and will return N^2 results for multiple entries with N occurrences.
Even simpler would be this:
select e.employee_ID, d.dep_name from employee e,Departement d where d.dep_id= e.departement_id
I have one table that holds my ressources:
Ressource | Ressource-ID
And a table that holds the associations
Ressource-ID | Employee-ID
How to select the ressources of an Employee that are available, i.e. not in the association table?
I've tried this, but it's not working:
select r.ress, r.ress_id
FROM Ressource r
LEFT outer JOIN Ressource_Employee_Association a ON r.ress_id = a.ress_id
WHERE a.emp_id = 'ID00163efea66b' and a.ress_id IS NULL
Any ideas?
Thanks
Thomas
After writing my above comments, and looking at the proposed solutions: I think I've got some more understanding of what you are trying to do.
Assuming you have unlimited quantity of resources in your resources table, you want to select the un-assigned resources per employee (based on their non-existence for any specific employee in the resource association table).
In order to accomplish this (and get a comprehensive list of employees) you'll need a 3rd table, in order to reference the complete list of employees. You'll also need to CROSS JOIN all of the resources onto the list of employees (assuming every employee has access to every resource), then you LEFT JOIN (LEFT OUTER JOIN whatever) your association list onto the query where the resource_id and employee_id match the resource_id in the resources table, and the employee_id in the employees table (respectively). THEN you add your where clause that filters out all the records that assign an employee to a resource. This leaves you with the resources that are available to the employee, which they also do not have signed out. This is convoluted to say, so hopefully the query sheds more light:
SELECT e.employee_id, e.employee, r.res_id, r.res
FROM employees e
CROSS JOIN resources r
LEFT JOIN assigned_resources ar
ON ar.employee_id = e.employee_id AND r.res_id = ar.res_id
WHERE ar.res_id IS NULL
If you don't have an employees table, you can accomplish the same by using the assigned resources table, but you will be limited to selecting employees who already have some resources allocated. You'll need to add a GROUP BY query because of the possible existence of multiple employee definitions in the association table. Here's what that query would look like:
SELECT e.employee_id, r.res_id, r.res
FROM assigned_resources e
CROSS JOIN resources r
LEFT JOIN assigned_resources ar
ON ar.employee_id = e.employee_id AND r.res_id = ar.res_id
WHERE ar.res_id IS NULL
GROUP BY e.employee_id, r.res_id
Does this work?
select r.ress, r.ress_id
from resource r
where not exists
(
select 1 from ressource_emplyee_association a
where a.emp_id = '...' and a.ress_id = r.ress_id
)
EDIT
Before that I had the following, but changed it according to the comments below:
select r.ress, r.ress_id
from resource r
where not exists
(
select top 1 1 from ressource_emplyee_association a
where a.emp_id = '...' and a.ress_id = r.ress_id
)
The WHERE clause is applied after the LEFT JOIN. This means that you are currently trying to get results where there is NO matching record in Ressource_Employee_Association, but where the emp_id equals 'ID00163efea66b'.
But if there is no matching record, how can emp_id be anything other than NULL?
One option is to move part of the WHERE clause into the join...
SELECT
r.ress, r.ress_id
FROM
Ressource r
LEFT OUTER JOIN
Ressource_Employee_Association a
ON r.ress_id = a.ress_id
AND a.emp_id = 'ID00163efea66b'
WHERE
a.ress_id IS NULL
This will list all resources that are not associated to employee 'ID00163efea66b'.
EDIT
Your comment implies that what you want is...
- A view listing all employees
- For each employee list each resource that they DON'T have
This requires an extra table listing all of your employees.
SELECT
*
FROM
Employee
CROSS JOIN
Ressource
WHERE
NOT EXISTS (SELECT * FROM Ressource_Employee_Association
WHERE emp_id = Employee.id
AND ress_id = Ressource.id)
SELECT *
FROM Ressource
WHERE ress_id IN (
SELECT ress_id,
FROM Ressource
MINUS
SELECT ress_id
FROM Ressource_Employee_Association
WHERE emp_id = 'ID00163efea66b'
);