ROW - NAME - BRAND
____________________
1 - John - Ford
2 - John - BMW
3 - John - Merc
4 - Mary - Ford
5 - Mary - VW
6 - James - Ford
7 - James - Jeep
8 - James - Lexus
9 - Susan - Jeep
10 - Susan - Lexus
How do I find the values in Column A that does not have a particular value in Column B.
Column A, let's say Name, is not unique, there are multiple rows with the same value in A (as the same person can have multiple cars)
Column B, let's say Car Brand, shows the brand of car that the person has in A. There are only a total of 5 brands possible and ideally everyone should have one of the brands - Ford.
How do I find out all the unique Names for people who have other car types, but are missing the Ford? In the above example I'm looking to find Susan.
You can use aggregation:
select name
from t
group by name
having sum(brand = 'Ford') = 0;
The having clause counts the number of rows that match Ford. The = 0 says there are no such rows.
I hope you are talking about this.
SELECT DISTINCT name
FROM tablename
WHERE brand <> 'FORD';
If you need names with all 4 other cars but with no Ford:
SELECT DISTINCT name FROM t
WHERE (SELECT COUNT(DISTINCT car) FROM t AS t2 WHERE t2.name=t.name)=4
AND (SELECT COUNT(*) FROM t AS t2 WHERE t2.name=t.name AND t2.car="Ford")=0
There are a lot of ways, but one answer would be:
SELECT DISTINCT A
FROM T
WHERE A NOT IN (
SELECT A
FROM T
WHERE B = 'Ford'
);
Related
I'm defining the relationship between the two tables using a join table. I want to arrange them in the order of many overlapping things. Currently, we are using subquery, is there a way to get the same result using join?
People FoodTable PeopleFood
ID | NAME ID | Food ID | PeopleId | FoodId
1 BOB 1 Hamberger 1 1 1
2 JOHN 2 Pizza 2 1 2
3 KATY 3 Chicken 3 1 3
4 MILLER 4 Salad 4 2 1
5 AMANDA 5 Sushi 5 2 2
6 2 3
7 3 2
8 3 3
9 4 3
10 4 5
11 5 5
When the table is defined in this way, I want to arrange food tastes similar to Bob's.
I'm doing it like this now.
SELECT people_id, COUNT(people_id) as count
FROM peopleFood
WHERE food_id IN
(SELECT food_id FROM peopleFood
WHERE people_id = 1)
AND people_id != 1
GROUP BY people_id
ORDER BY count DESC;
-- Result -------------
People_id | count
2 3
3 2
4 1
Is there a better way to change this method or use join?
Thank you!!!
You have been inconsistent in your use of the table and column names -
Tables - PeopleFood in your sample data but you reference peopleFood in your query.
Columns - PeopleId and FoodId in your sample data but you reference people_id and food_id in your query.
Choose a naming convention and stick to it. Everyone has there own preference but the important thing is to be consistent.
The equivalent query with INNER JOIN instead of your sub-query is -
SELECT
`pf2`.`people_id`,
COUNT(`pf2`.`food_id`) as `count`
FROM `PeopleFood` `pf1`
INNER JOIN `PeopleFood` `pf2`
ON `pf2`.`people_id` <> `pf1`.`people_id`
AND `pf2`.`food_id` = `pf1`.`food_id`
WHERE `pf1`.`people_id` = 1
GROUP BY `pf2`.`people_id`
ORDER BY `count` DESC;
The performance difference between the two queries is unlikely to be noticeable and it might be argued that the intent is clearer in your version with the sub-query.
The surrogate key ID on your PeopleFood table should be dropped in favour of the compound “natural” primary key on people_id and food_id.
The Cost of Useless Surrogate Keys in Relationship Tables
Inner join:
SELECT p.People_id, COUNT(p.People_id) as count FROM PeopleTable p
INNER JOIN FoodTable f
ON(p.People_id = f.FoodId)
WHERE people = 1
GROUP BY p.people_id
ORDER BY count DESC;
If it helps, please mark it as an accepted answer!
In MySQL I have a table.
Example:
id name type
1 Thomas 2
2 Thomas 2
3 Thomas 1
4 Paul 3
5 Paul 4
6 Paul 4
I need calculate same records by 2 columns.
Result for this example should be:
name type countOfRecords
Thomas 2 2
Thomas 1 1
Paul 3 1
Paul 4 2
Could you help me with this request?
Since you want records in your result set for each name and type <name,type> pair
you need to group by name and type.
SELECT
name,
type,
COUNT(*) countOfRecords
FROM your_table
GROUP BY name,type;
Note:
Group BY <some column> would generate a result set where number of rows = number of distinct / different / unique <some column>.
Same holds for multiple columns in GROUP BY clause.
This may be right solution:
SELECT name, type, COUNT(*) as countOfRecords
FROM your_table
GROUP BY name,type;
I am having some major difficulties with grouping and summing results of a query in the way I want. I am having trouble explaining what I want to do in words, so I'll just show you. I have three tables: A, H, and W which look like (simplified):
Table A:
Aid name
1 adam
2 bob
Table H:
Hid Wid date atk Aid
1 1 - 10 2
2 2 - 1 1
3 2 - 5 1
4 1 - 2 2
5 1 - 22 1
6 2 - 7 2
Table W:
Wid name user pass
1 charlie - -
2 donald - -
I am trying to get the SUM of atk grouped by Aid and Wid. Basically, assume this is a fight club tally. I want to display the sum of how many times person W attacked person A (it will always be a one directional fight, ie: charlie can only attack adam, but adam can't attack charlie). (not really a fight club - being used for an online game :))
I am trying to get my result to look like:
name1 atk name2
charlie 22 adam
charlie 12 adam
donald 6 bob
donald 7 bob
My current query looks like...
SELECT w.name AS name1, h.atk, a.name AS name2
FROM H
JOIN W ON w.Wid=h.Wid
JOIN A ON a.Aid=h.Aid
...which gives me every instance that name1 attacked name2. When I try to GROUP BY and/or SUM(h.atk) it is grouping or summing in a way I can't figure out. I'm just not understanding how to accomplish this. Any help would be greatly appreciated.
Thanks in advance!
SELECT w.name AS name1, sum(h.atk) as atk, a.name AS name2
FROM H
JOIN W ON w.Wid=h.Wid
JOIN A ON a.Aid=h.Aid
GROUP BY w.name, a.name
Let's say I have the following table :
Name - Country - Age
--------------------
Toto - Switzerland - 10
Titi - France - 12
Tata - Italy - 21
Tutu - England - 13
Tete - Italy - 14
I want to create a sql query as simple as possible to regroup people living in defined grouped countries like :
Group A = Switzerland + Italy
Group B = France + England
I don't know how to create a group withn my records with a column that could have multiple different values in the same group...
Could somebody help me with this ?
More information : SQL Server 2008 database.
You mean like this?
SELECT COUNT(Name), GroupA, GroupB FROM
(`SELECT Name, Country, Age,
Country='Switzerland' OR Country='Italy' As GroupA,
Country='France' OR Country='England' As GroupB)
Group By GroupA, GroupB
Select * from (select *,case when Country ='Switzerland' then 'A'
when Country ='Italy' then 'A'
when Country ='France' then 'B'
when Country ='England' then 'B'
else 'C' end) classification from table1)
order by classification
This will group the ppl as per your criteria. If this grouping is static you can have seprate table and use inner join. That will make query more readable
Whilst trying to do pagination I've run into this problem.
My table-
ID CarBrand Car Model
---------------------------
1 Alfa Romeo Guilietta
2 Alfa Romeo Mito
3 Audi A3
4 Audi R8
5 Audi TT
6 Fiat Punto
7 Fiat Panda
8 Ford Mondeo
9 Ford Mustang
10 Nissan Almera
11 Nissan Note
12 Nissan Qashqai
13 Toyota Aygo
14 Toyota Prius
15 Volkswagen Beetle
16 Volkswagen Golf
17 Volkswagen Polo
18 Volkswagen Up
I have the data displayed like so, in groups of two:
-Fiat - Punto
Panda
-Ford - Mondeo
Mustang
So there are 2 brands but 4 database results.
Is it possible to have a query limit and offset my results to two brands while showing all the models for the brand?
Sorry if I'm not clear!
It is clear. Try this:
select * from t t1
join (
select distinct carBrand from t
limit 2
) s on t1.carBrand = s.carBrand
Before the limit 2 apply the ordering you want.
To get a limit, without using the limit keyword, you can impose a count.
For example, given the table definition
create table cars (id int,
carBrand char(10),
carModel char(10));
this will give you all the Car Models for the top 2 Car Brands
select cars.carBrand, cars.carModel
from cars
where ((select count(*) from
(select distinct carBrand from cars) as carBrands
where carBrands.carBrand < cars.carBrand) < 2)
order by cars.carBrand, cars.carModel;
This creates an inline table just listing the carBrands and then joins this back to cars to get the list of all cars that are in the top 2 brands. The count(*) .... < 2 enforces the limit. Consider 'Ford', for example, in your above data. In 'Ford''s case, there are 3 brands that are < 'Ford' alphabetically, so the count(*) above = 3. Since 3 is not less than 2, no 'Ford' cars appear in the output.
The output on your test data would be:
CARBRAND CARMODEL
Alfa Romeo Guilietta
Alfa Romeo Mito
Audi A3
Audi R8
Audi TT
Now, you didn't say how you wanted to pick the 2 brands -- you just listed Ford and Fiat in your example -- I don't know how you happened to pick those. If you want something other than alphabetical criteria for ordering, that's doable, but harder.
SQL Fiddle and results for all this: http://sqlfiddle.com/#!2/33a8f/3
It's a matter of database design. Maybe you should split your data into two tables model (model names) and brand (brand names). Then you can write a query like this:
SELECT m.name, b.name
FROM model m
INNER JOIN brand b
WHERE b.id IN (
SELECT id
FROM brand
ORDER BY name ASC
LIMIT 2 OFFSET 0
)
I did not test the code. No need for GROUP BY in my opinion.