I have two tables, one with persons and the other with various dogs that each person owns. I would like to only query the database once.
I would like to always display the active people, even if they don't have any dogs attached to their id. How would I query this?
SELECT person.*, dog.*
FROM person
LEFT OUTER JOIN dog ON person.id = dog.person_id AND
WHERE person.active=true and dog.active=true
Move the condition to the join (on clause); that way you will left join any records that match both the person_id and the active flag. If no active dog is found, the person is still present in the result.
If you add the condition in the where clause, first the left join is evaluated, and then all records that are not active dogs (including those that don't have a dog at all) are filtered out. It effectively turns your left join into an inner join if you do that.
SELECT
person.*,
dog.*
FROM
person
LEFT JOIN dog ON person.id = dog.person_id AND dog.active=true
WHERE
person.active=true
Related
I want to show number of employees in each role in each department.For this I have created below mentioned query. But it does not give (a) ALL the roles (b) All the employees in role in respective department where employee count is zero.
SELECT qodi.department_name,IFNULL(qobi.band_name, 'No Band') 'Band',count(qu1.user_id) emp_count
FROM
qee_org_dept_info qodi
LEFT OUTER JOIN qee_user qu1 ON qodi.department_name =qu1.department AND qu1.org_id=1
LEFT OUTER JOIN qee_org_band_info qobi ON qobi.band_name = qu1.band_name
GROUP BY qodi.department_name,qobi.band_name
ORDER BY qodi.department_name
Roles are defined in qee_org_band_info. Users are defined in qee_user and departments in qee_org_dept_info
Result of above query is:-
Expected Results is :- All departments and all roles should come, even if employee count for that role in respective department is zero.
All departments and all roles should come, even if employee count for that role in respective department is zero.
This means, that you need to produce a cartesian join between the departments and roles and join the employees on the result of the cartesian join.
SELECT qodi.department_name,IFNULL(qobi.band_name, 'No Band') 'Band',count(qu1.user_id) emp_count
FROM
(qee_org_dept_info qodi INNER JOIN qee_org_band_info qobi)
LEFT OUTER JOIN qee_user qu1 ON qodi.department_name =qu1.department AND qobi.band_name = qu1.band_name AND qu1.org_id=1
GROUP BY qodi.department_name,qobi.band_name
ORDER BY qodi.department_name
Pls note that I changed the order of the joins and that I used inner join instead of left to join departments with bands. According to mysql documentation:
INNER JOIN and , (comma) are semantically equivalent in the absence of a join condition: both produce a Cartesian product between the specified tables (that is, each and every row in the first table is joined to each and every row in the second table).
Technically, there is no need to have parentheses around the inner join expression, however, I feel that the code is more readable this way.
I'm not entirely sure that what I'm trying to do is possible. Can you use an OR in the condition of a left join? I start from my users table and then it can either go from week_meal to meal (adding a meal they do not own to their weekly meal plan) or straight to meal (a meal they own). That part appears to be working, but when I include mta.meal_to_add_id in the select, it incorrectly pulls in meals that do NOT meet the criteria in the LEFT JOIN to meal_to_add.
Fiddle with structure: http://sqlfiddle.com/#!2/7bd9c4
SELECT DISTINCT m.*, o.username as owner,i.*, mta.meal_to_add_id, follow_id
FROM webusers wu
LEFT JOIN week_meal wm ON wu.id=wm.user_id
LEFT JOIN meal m ON (wu.id=m.user_id OR wm.meal_id=m.meal_id)
LEFT JOIN webusers o ON m.user_id=o.id
LEFT JOIN meal_to_add mta ON
((wm.user_id = mta.user_id AND wm.meal_id=mta.meal_id)
OR (m.user_id=mta.user_id AND m.meal_id=mta.meal_id))
JOIN ingredient i ON m.meal_id = i.meal_id
LEFT JOIN follow f ON
m.user_id!=wu.id AND
m.user_id=f.followed_webuser_id
AND wu.id=f.followee_webuser_id
WHERE wu.id=5 AND m.meal_id in (138)
ORDER BY m.meal, i.ingredient_id
OUTPUT: It should be just like this only including the field for mta.meal_to_add_id, which in this case should be NULL for all rows (18)
Sample Results
To answer the first part of your question: Yes, you can use an OR clause in a LEFT JOIN.
As for the second part, in plain words, this is what the query seems to say:
Join 'week meals' to users on the user id. Join meals to users on that same user id OR join meals to users on the meal id. Assume now that we have some matching meal/user combinations, where some meal rows are matched on user, and others are matched on the meal id.
Next, join webusers to meals again. Now we have meal rows possibly matching two sets of users. So when mta tries to match meals, it is matching two possible sets of meal/user combinations.
My practice in cases like this is to break up the query into two queries and put the intermediate results in a temp table (using MEMORY engine), then select from that.
I'm trying to make something like a chart list and my actual query is like this:
SELECT
user.username,
songs.*,
albums.desc,
albums.release,
albums.name,
AVG(songrating.rating)
FROM
songs
INNER JOIN
user
ON
songs.userid=user.id
INNER JOIN
albums
ON
songs.albumid=albums.id
INNER JOIN
songrating
ON
songs.id=songrating.songid
GROUP BY
songrating.songid
it only shows entries with at least one rating entry, but I also want these without a rating
I tried to use it with if / case but it doesn't work (or I'm doing something wrong)
If I remove the avg etc, it works correctly.
It may be JOIN issue
The INNER JOIN keyword selects all rows from both tables as long as there is a match between the columns in both tables.
The LEFT JOIN keyword returns all rows from the left table (table1), with the matching rows in the right table (table2). The result is NULL in the right side when there is no match.
Try LEFT JOIN
I have these tables
AssigenmentList --linksto ---School,AgeGroup
Users will have birthday attached to it
AgeGroup in turn linked to many ages like AgeGroup 3-4 is linked to one to many with 3,4 in numerical format
Now i want all the Assignment list which are linked to particular SCHOOL belong to same age as the age of Child
As a general rule:
select a.*, b.*, c.* from
A a inner join B b on a.idB = b.id
inner join C c on b.idC = c.id
You use inner join if a.idB must have match to add the row to the resultset. Left outer join if the mere presence of a.idB (left side) is enough to project the row.
The trick is to navigate from the starting table to the last joining the columns that ties them.
I don't understand the concept of a left outer join, a right outer join, or indeed why we need to use a join at all! The question I am struggling with and the table I am working from is here: Link
Question 3(b)
Construct a command in SQL to solve the following query, explaining why it had to employ the
(outer) join method. [5 Marks]
“Find the name of each staff member and his/her dependent spouse, if any”
Question 3(c) -
Construct a command in SQL to solve the following query, using (i) the join method, and (ii) the
subquery method. [10 Marks]
“Find the identity name of each staff member who has worked more than 20 hours on the
Computerization Project”
Can anyone please explain this to me simply?
Joins are used to combine two related tables together.
In your example, you can combine the Employee table and the Department table, like so:
SELECT FNAME, LNAME, DNAME
FROM
EMPLOYEE INNER JOIN DEPARTMENT ON EMPLOYEE.DNO=DEPARTMENT.DNUMBER
This would result in a recordset like:
FNAME LNAME DNAME
----- ----- -----
John Smith Research
John Doe Administration
I used an INNER JOIN above. INNER JOINs combine two tables so that only records with matches in both tables are displayed, and they are joined in this case, on the department number (field DNO in Employee, DNUMBER in Department table).
LEFT JOINs allow you to combine two tables when you have records in the first table but might not have records in the second table. For example, let's say you want a list of all the employees, plus any dependents:
SELECT EMPLOYEE.FNAME as employee_first, EMPLOYEE.LNAME as employee_last, DEPENDENT.FNAME as dependent_last, DEPENDENT.LNAME as dependent_last
FROM
EMPLOYEE INNER JOIN DEPENDENT ON EMPLOYEE.SSN=DEPENDENT.ESSN
The problem here is that if an employee doesn't have a dependent, then their record won't show up at all -- because there's no matching record in the DEPENDENT table.
So, you use a left join which keeps all the data on the "left" (i.e. the first table) and pulls in any matching data on the "right" (the second table):
SELECT EMPLOYEE.FNAME as employee_first, EMPLOYEE.LNAME as employee_last, DEPENDENT.FNAME as dependent_first, DEPENDENT.LNAME as dependent_last
FROM
EMPLOYEE LEFT JOIN DEPENDENT ON EMPLOYEE.SSN=DEPENDENT.ESSN
Now we get all of the employee records. If there is no matching dependent(s) for a given employee, the dependent_first and dependent_last fields will be null.
example (not using your example tables :-)
I have a car rental company.
Table car
id: integer primary key autoincrement
licence_plate: varchar
purchase_date: date
Table customer
id: integer primary key autoincrement
name: varchar
Table rental
id: integer primary key autoincrement
car_id: integer
bike_id: integer
customer_id: integer
rental_date: date
Simple right? I have 10 records for cars because I have 10 cars.
I've been running this business for 10 years, so I've got 1000 customers.
And I rent the cars about 20x per year per cars = 10 years x 10 cars x 20 = 2000 rentals.
If I store everything in one big table I've got 10x1000x2000 = 20 million records.
If I store it in 3 tables I've got 10+1000+2000 = 3010 records.
That's 3 orders of magnitude, so that's why I use 3 tables.
But because I use 3 tables (to save space and time) I have to use joins in order to get the data out again
(at least if I want names and licence plates instead of numbers).
Using inner joins
All rentals for customer 345?
SELECT * FROM customer
INNER JOIN rental on (rental.customer_id = customer.id)
INNER JOIN car on (car.id = rental.car_id)
WHERE customer.id = 345.
That's an INNER JOIN, because we only want to know about cars linked to rentals linked to customers that actually happened.
Notice that we also have a bike_id, linking to the bike table, which is pretty similar to the car table but different.
How would we get all bike + car rentals for customer 345.
We can try and do this
SELECT * FROM customer
INNER JOIN rental on (rental.customer_id = customer.id)
INNER JOIN car on (car.id = rental.car_id)
INNER JOIN bike on (bike.id = rental.bike_id)
WHERE customer.id = 345.
But that will give an empty set!!
This is because a rental can either be a bike_rental OR a car_rental, but not both at the same time.
And the non-working inner join query will only give results for all rentals where we rent out both a bike and a car in the same transaction.
We are trying to get and boolean OR relationship using a boolean AND join.
Using outer joins
In order to solve this we need an outer join.
Let's solve it with left join
SELECT * FROM customer
INNER JOIN rental on (rental.customer_id = customer.id) <<-- link always
LEFT JOIN car on (car.id = rental.car_id) <<-- link half of the time
LEFT JOIN bike on (bike.id = rental.bike_id) <<-- link (other) 0.5 of the time.
WHERE customer.id = 345.
Look at it this way. An inner join is an AND and a left join is a OR as in the following pseudocode:
if a=1 AND a=2 then {this is always false, no result}
if a=1 OR a=2 then {this might be true or not}
If you create the tables and run the query you can see the result.
on terminology
A left join is the same as a left outer join.
A join with no extra prefixes is an inner join
There's also a full outer join. In 25 years of programming I've never used that.
Why Left join
Well there's two tables involved. In the example we linked
customer to rental with an inner join, in an inner join both tables must link so there is no difference between the left:customer table and the right:rental table.
The next link was a left join between left:rental and right:car. On the left side all rows must link and the right side they don't have to. This is why it's a left join
You use outer joins when you need all of the results from one of the join tables, whether there is a matching row in the other table or not.
I think Question 3(b) is confusing because its entire premise wrong: you don't have to use an outer join to "solve the query" e.g. consider this (following the style of syntax in the exam paper is probably wise):
SELECT FNAME, LNAME, DEPENDENT_NAME
FROM EMPLOYEE, DEPENDENT
WHERE SSN = ESSN
AND RELATIONSHIP = 'SPOUSE'
UNION
SELECT FNAME, LNAME, NULL
FROM EMPLOYEE
EXCEPT
SELECT FNAME, LNAME, DEPENDENT_NAME
FROM EMPLOYEE, DEPENDENT
WHERE SSN = ESSN
AND RELATIONSHIP = 'SPOUSE'
In general:
JOIN joints two tables together.
Use INNER JOIN when you wanna "look up", like look up detailed information of any specific column.
Use OUTER JOIN when you wanna "demonstrate", like list all the info of the 2 tables.