(I don't think I have titled this question correctly - but I don't know how to describe it)
Here is what I am trying to do:
Let's say I have a Person table that has a PersonID field. And let's say that a Person can belong to many Groups. So there is a Group table with a GroupID field and a GroupMembership table that is a many-to-many join between the two tables and the GroupMembership table has a PersonID field and a GroupID field. So far, it is a simple many to many join.
Given a list of GroupIDs I would like to be able to write a query that returns all of the people that are in ALL of those groups (not any one of those groups). And the query should be able to handle any number of GroupIDs. I would like to avoid dynamic SQL.
Is there some simple way of doing this that I am missing?
Thanks,
Corey
select person_id, count(*) from groupmembership
where group_id in ([your list of group ids])
group by person_id
having count(*) = [size of your list of group ids]
Edited: thank you dotjoe!
Basically you are looking for Persons for whom there is no group he is not a member of, so
select *
from Person p
where not exists (
select 1
from Group g
where not exists (
select 1
from GroupMembership gm
where gm.PersonID = p.ID
and gm.GroupID = g.ID
)
)
You're basically not going to avoid "dynamic" SQL in the sense of dynamically generating the query at query time. There's no way to hand a list around in SQL (well, there is, table variables, but getting them into the system from C# is either impossible (2005 & below) or else annoying (2008)).
One way that you could do it with multiple queries is to insert your list into a work table (probably a process-keyed table) and join against that table. The only other option would be to use a dynamic query such as the ones specified by Jonathan and hongliang.
Related
I'm completely new to SQL and have a question. I am using is PostgreSQL.
I have two tables called "employees" and "offices"
The table "employees" have a list of unique employees with each having an OfficeID (The office where they work).
What I want to do is to "count" the number of appearances of the Office_ID and take that count into the table "offices" where the "office_ID" have a column called "number_of_employees".
Being completely new to SQL the only thing I have managed to even come close to this is fore example.
SELECT COUNT(*)
FROM employees
WHERE office_id = 203
But this only selects and gives the sum of rows with the id "203" that has to be manually entered into "number of empolyees"
What I want is a trigger function that updates the field "number_of_empolyees" when a new record is inserted into the table "empolyees"
A view is the way to go here.
I am assuming since you're completely new to SQL, you're unsure how to make it work (Edit: just seen your comment after posting :^D) .
The correct way to count employees for each office is:
SELECT office_id, COUNT(*) as employeeCount
FROM employees
GROUP BY office_id
Note how your WHERE office_id = XXX has been replaced by a GROUP BY office_id in order to count employees for all offices in a single query.
That being done, we can use it inside the view.
Be careful about the JOIN: I believe in your schema, an office may have no employee (for instance, right after you created it or right before you delete it). We will handle that part with a LEFT JOIN.
CREATE VIEW OfficeWithEmployeeCount AS
SELECT Offices.*, EmployeeCount
FROM Offices
LEFT OUTER JOIN (SELECT office_id, COUNT(*) as EmployeeCount FROM Employees GROUP BY office_id) T
ON Offices.office_id = T.office_id
Note: to avoid having NULL returned in EmployeeCount for empty offices, you may want to write:
CREATE VIEW OfficeWithEmployeeCount AS
SELECT Offices.*, COALESCE(EmployeeCount,0)
FROM ...
Is there a way to tell Talend not to remove the prefix of column names especially when they are specified in the query to retrieve data from data source and keep the names mentioned in the query itself?
Thanks!
Assuming you are using the 'guess schema' feature with a query that joins some tables. Further assuming your tables have columns with the same names you run into trouble with the guessed schema. There is no way to have talend use or even know the names of the tables the colums come from, because they are part of a 'projection' and could result from transformation and/or aggregation. Thus, you'll need to help talend guessing the correct schema, which means a) you cant use the * to select all columns and b) you should assign each column an alias that hints at the table the column comes from.
So instead of select * from employee join department on employee.department_id = department.id you'd have something like select e.id as emp_id, e.name as emp_name, d.id as department_id, d.name as department_name from employee e join department d on e.department_id = d.id. The id from employee will be emp_id in the guessed schema.
Okay. So the question that I got asked by the teacher was this:
(5 marks) Construct a SQL query on the dvdrental database that uses a natural join of two or more tables and an additional where condition. (E.g. find the titles of films rented by a particular customer.) Note the hints on the course news page if your query returns nothing.
Here is the layout of the database im working with:
http://www.postgresqltutorial.com/wp-content/uploads/2013/05/PostgreSQL-Sample-Database.png
The hint to us was this:
PostgreSQL hint:
If a natural join doesn't produce any results in the dvdrental DB, it is because many tables have the last update: timestamp field, and thus the natural join tries to join on that field as well as the intended field.
e.g.
select *
from film natural join inventory;
does not work because of this - it produces an empty table (no results).
Instead, use
select *
from film, inventory
where film.film_id = inventory.film_id;
This is what I did:
select *
from film, customer
where film.film_id = customer.customer_id;
The problem is I cannot get a particular customer.
I tried doing customer_id = 2; but it returns a error.
Really need help!
Well, it seems that you would like to join two tables that have no direct relation with each other, there's your issue:
where film.film_id = customer.customer_id
To find which films are rented by which customer you would have to join customer table with rental, then with inventory and finally with film.
The task description states
Construct a SQL query on the dvdrental database that uses a natural join of two or more tables and an additional where condition.quote
I have created a database as part of university assignment and I have hit a snag with the question in the title.
More likely I am being asked to find out how many films each company has made. Which suggests to me a group by query. But I have no idea where to begin. It is only a two mark question but the syntax is not clicking in my head.
My schema is:
CREATE TABLE Movie
(movieID CHAR(3) ,
title CHAR(36),
year NUMBER,
company CHAR(50),
totalNoms NUMBER,
awardsWon NUMBER,
DVDPrice NUMBER(5,2),
discountPrice NUMBER(5,2))
There are other tables but at first glance I don't think they are relevant to this question.
I am using sqlplus10
The answer you need comes from three basic SQL concepts, I'll step through them with you. If you need more assistance to create an answer from these hints, let me know and I can try to keep guiding you.
Group By
As you mentioned, SQL offers a GROUP BY function that can help you.
A SQL Query utilizing GROUP BY would look like the following.
SELECT list, fields, aggregate(value)
FROM tablename
--WHERE goes here, if you need to restrict your result set
GROUP BY list, fields
a GROUP BY query can only return fields listed in the group by statement, or aggregate functions acting on each group.
Aggregate Functions
Your homework question also needs an Aggregate function called Count. This is used to count the results returned. A simple query like the following returns the count of all records returned.
SELECT Count(*)
FROM tablename
The two can be combined, allowing you to get the Count of each group in the following way.
SELECT list, fields, count(*)
FROM tablename
GROUP BY list, fields
Column Aliases
Another answer also tried to introduce you to SQL column aliases, but they did not use SQLPLUS syntax.
SELECT Count(*) as count
...
SQLPLUS column alias syntax is shown below.
SELECT Count(*) "count"
...
I'm not going to provide you the SQL, but instead a way to think about it.
What you want to do is select where the company matches and count the total rows returned. That count is the number of films made by the specified company.
Hope that points you in the right direction.
Select company, count(*) AS count
from Movie
group by company
select * group by company won't work in Oracle.
I'm copying data from one database to another and massaging the data while I'm at it. Both databases have tables called Clients and Jobs.
However, in database "Alpha" the Jobs table does not have a relationship to the Clients table, where database "Epsilon" does. Alpha's Jobs table just has the Clients name in an nvarchar column.
I need a select statement to lookup the Client's ID in the Client table by their name while I am inserting it into the Jobs table in Epsilon.
My unfinished SQL statement looks like this:
insert into Epsilon.dbo.Jobs (ClientId, Name, Location, DateCreated)
select ????, Name, Location, DateCreated from Alpha.dbo.Jobs
How can I modify this so that the ???? contains the ClientId from the Clients table in Epsilon? I know I need to lookup the data using the Name column in Jobs, but I can't figure out the syntax for this.
What you need is a join. Joins, contrary to what pretty much everybody thinks when starting out, don't require defined relationships in the schema of the database. They just require that the two columns you're comparing have the same type (edit see comments).
The question is which join do you want. Because there isn't a relationship defined, there may be clients that have jobs and clients that don't, and jobs that have clients and jobs that don't.
I'm assuming that you want all JOBS that exist, and where a ClientId matches the CLIENTS table bring in the ClientId, and where that relationship doesn't exist to leave the ClientId null. We can do this with a LEFT JOIN. Jobs LEFT JOIN Clients will bring in all records on the LEFT, even where the relationship defined with Clients on the right doesn't exist. We could reverse the two and do a RIGHT JOIN, but that's not what people usually do. I'll leave it to you to read up on other types of joins and how they work.
So your select statement would look like:
select ClientId, Name, Location, DateCreated
from Alpha.dbo.Jobs as J LEFT JOIN
Alpha.dbo.Clients as C ON j.ClientName = c.ClientName
If Jobs.ClientName is not the same data type as c.ClientName, you can edit the schema before running the query to bring them in line with each other.
insert into Epsilon.dbo.Jobs (ClientId, Name, Location, DateCreated)
select c.ClientID, a.Name, a.Location, a.DateCreated from Alpha.dbo.Jobs a
join Epsilon.dbo.Client c on c.Name = a.ClientName
This is a pretty optimistic join, but even if it needs to be modified this should give you the general idea.
insert into Epsilon.dbo.Jobs
(ClientId, Name, Location,
DateCreated)
select c.ClientId, Name, Location, DateCreated from Alpha.dbo.Jobs as j
inner join Epsilon.dbo.Clients as c On
(j.ClientId = c.ClientId)