I have a large dataset in Microsoft Access with postcodes in one column and an unordered category in the next as follows:
Postcode | Class
--------------------------
1111AA | A
1111AA | B
1111AA | A
1111AB | C
1111AB | C
1111AB | A
I would like to group the data such that on the left-hand side I have one Postcode for the mode of the Class on the right. The classes are unordered (i.e: A is not better than B, nor C better than B). I have tried using queries but they only really work for numerical data and I can only seem to use these techniques for finding averages.
So in the end I want:
Postcode | Class
------------------------
1111AA | A
1111AB | C
via most frequency per postal code, you would need a pre-query first and then match that count... basically doing a triple-process against the table.
The third level in alias (QPerClass) gets on a per postal code, each class and their respective counts. A max classification count of 3 in one postal code vs count 5 in another are two separate things, and you don't want the overall most popular 5 count if such other postal code only has 3, it would never find a corresponding match.
So, per your sample data, this would result in
PostCode Class Count
1111AA A 2
1111AA B 1
1111AB C 2
1111AB A 1
From that, we need each postal code's MAXIMUM COUNT, but you can't grab the class associated with it as which you can't say grab the class with the max count, and you cant do a limit 1 per postal code.. this would result in
1111AA 2
1111AB 2
Now that you have the count per postal code, join that to your original table and apply a group by AND a having so that the outer-level HAVING COUNT(*) matches the MAX() count as determined in the second step.
select
pc.postCode,
pc.class,
MaxPostByCode.MaxCount
from
PostalCodes pc
JOIN ( select QPerClass.postCode,
max( QPerClass.perClassCount ) MaxCount
from
( select pc2.postcode,
pc2.class,
count(*) perclassCount
from
postalcodes pc2
group by
pc2.postcode,
pc2.class ) QPerClass
group by
QPerClass.postCode ) MaxByPostCode
on pc.postcode = MaxByPostCode.postCode
group by
pc.postCode,
pc.class,
MaxPostByCode.MaxCount
having
count(*) = MaxPostByCode.MaxCount
Now, if you have an instance where there are multiple entries that have the same MODE (max count per class), then you would have to wrap it up yet again to get the MIN( CLASS ) that qualified the HAVING clause grouped by postal code, such as
select
m.postcode,
min( m.class )
from
( entire query above with the HAVING clause ) m
group by
m.postcode
This is what you want...
SELECT TOP 1 WITH TIES PostCode, Class
FROM #Temp
GROUP BY PostCode, Class
ORDER BY COUNT(*) DESC
Use a correlated subquery to get the most often occurring class for a postcode:
select postcode, class
from mytable as m
where class =
(
select top 1 m2.class
from mytable m2
where m2.postcode = m.postcode
group by m2.class
order by count(*) desc
)
group by m.postcode, m.class;
In case of a tie, one class is picked arbitrarily.
Related
Let's say if I have a table that contains Equipment IDs of equipments for each Equipment Type and Equipment Age, how can I do a Count Distinct of Equipment IDs that have at least that Equipment Age.
For example, let's say this is all the data we have:
equipment_type
equipment_id
equipment_age
Screwdriver
A123
1
Screwdriver
A234
2
Screwdriver
A345
2
Screwdriver
A456
2
Screwdriver
A567
3
I would like the output to be:
equipment_type
equipment_age
count_of_equipment_at_least_this_age
Screwdriver
1
5
Screwdriver
2
4
Screwdriver
3
1
Reason is there are 5 screwdrivers that are at least 1 day old, 4 screwdrivers at least 2 days old and only 1 screwdriver at least 3 days old.
So far I was only able to do count of equipments that falls within each equipment_age (like this query shown below), but not "at least that equipment_age".
SELECT
equipment_type,
equipment_age,
COUNT(DISTINCT equipment_id) as count_of_equipments
FROM equipment_table
GROUP BY 1, 2
Consider below join-less solution
select distinct
equipment_type,
equipment_age,
count(*) over equipment_at_least_this_age as count_of_equipment_at_least_this_age
from equipment_table
window equipment_at_least_this_age as (
partition by equipment_type
order by equipment_age
range between current row and unbounded following
)
if applied to sample data in your question - output is
Use a self join approach:
SELECT
e1.equipment_type,
e1.equipment_age,
COUNT(*) AS count_of_equipments
FROM equipment_table e1
INNER JOIN equipment_table e2
ON e2.equipment_type = e1.equipment_type AND
e2.equipment_age >= e1.equipment_age
GROUP BY 1, 2
ORDER BY 1, 2;
GROUP BY restricts the scope of COUNT to the rows in the group, i.e. it will not let you reach other rows (rows with equipment_age greater than that of the current group). So you need a subquery or windowing functions to get those. One way:
SELECT
equipment_type,
equipment_age,
(Select COUNT(*)
from equipment_table cnt
where cnt.equipment_type = a.equipment_type
AND cnt.equipment_age >= a.equipment_age
) as count_of_equipments
FROM equipment_table a
GROUP BY 1, 2, 3
I am not sure if your environment supports this syntax, though. If not, let us know we will find another way.
I would like to ignore some of the results of my query as for all intents and purposes, some of the results are a duplicate, but based on the way the request was made, we need to use this hierarchy and although we are seeing different 'Company_Name' 's, we need to ignore one of the results.
Query:
SELECT
COUNT(DISTINCT A12.Company_name) AS Customer_Name_Count,
Company_Name,
SUM(Total_Sales) AS Total_Sales
FROM
some_table AS A12
GROUP BY
2
ORDER BY
3 ASC, 2 ASC
This code omits half a doze joins and where statements that are not germane to this question.
Results:
Customer_Name_Count Company_Name Total_Sales
-------------------------------------------------------------
1 3 Blockbuster 1,000
2 6 Jimmy's Bar 1,500
3 6 Jimmy's Restaurant 1,500
4 9 Impala Hotel 2,000
5 12 Sports Drink 2,500
In the above set, we can see that numbers 2 & 3 have the same count and the same total_sales number and similar company names. Is there a way to create a case statement that takes these 3 factors into consideration and then drops one or the other for Jimmy's enterprises? The other issue is that this has to be variable as there are other instances where this happens. And I would only want this to happen if the count and sales number match each other with a similar name in the company name.
Desired result:
Customer_Name_Count Company_Name Total_Sales
--------------------------------------------------------------
1 3 Blockbuster 1,000
2 6 Jimmy's Bar 1,500
3 9 Impala Hotel 2,000
4 12 Sports Drink 2,500
Looks like other answers are accurate based on assumption that Company_IDs are the same for both.
If Company_IDs are different for both Jimmy's Bar and Jimmy's Restaurant then you can use something like this. I suggest you get functional users involved and do some data clean-up else you'll be maintaining this every time this issue arise:
SELECT
COUNT(DISTINCT CASE
WHEN A12.Company_Name = 'Name2' THEN 'Name1'
ELSE A12.Company_Name
END) AS Customer_Name_Count
,CASE
WHEN A12.Company_Name = 'Name2' THEN 'Name1'
ELSE A12.Company_Name
END AS Company_Name
,SUM(A12.Total_Sales) AS Total_Sales
FROM some_table er
GROUP BY CASE
WHEN A12.Company_Name = 'Name2' THEN 'Name1'
ELSE A12.Company_Name
END
Your problem is that the joins you are using are multiplying the number of rows. Somewhere along the way, multiple names are associated with exactly the same entity (which is why the numbers are the same). You can fix this by aggregating by the right id:
SELECT COUNT(DISTINCT A12.Company_name) AS Customer_Name_Count,
MAX(Company_Name) as Company_Name,
SUM(Total_Sales) AS Total_Sales
FROM some_table AS A12
GROUP BY Company_id -- I'm guessing the column is something like this
ORDER BY 3 ASC, 2 ASC;
This might actually overstate the sales (I don't know). Better would be fixing the join so it only returned one name. One possibility is that it is a type-2 dimension, meaning that there is a time component for values that change over time. You may need to restrict the join to a single time period.
You need to have function to return a common name for the companies and then use DISTINCT:
SELECT DISTINCT
Customer_Name_Count,
dbo.GetCommonName(Company_Name) as Company_Name,
Total_Sales
FROM dbo.theTable
You can try to use ROW_NUMBER with window function to make row number by Customer_Name_Count and Total_Sales then get rn = 1
SELECT * FROM (
SELECT *,ROW_NUMBER() OVER(PARTITION BY Customer_Name_Count,Total_Sales ORDER BY Company_Name) rn
FROM (
SELECT
COUNT(DISTINCT A12.Company_name) AS Customer_Name_Count,
Company_Name,
SUM(Total_Sales) AS Total_Sales
FROM
some_table AS A12
GROUP BY
Company_Name
)t1
)t1
WHERE rn = 1
Note: The Data schema can not be changed. I'm stuck with it.
Database: SQLite
I have a simple tree structure, without parent keys, that is only 1 level deep. I have simplied the data for clarity:
ID Content Title
1 Null Canada
2 25 Toronto
3 33 Vancouver
4 Null USA
5 45 New York
6 56 Dallas
The structure is ordinal as well so all Canadian Cities are > Canada's ID of 1 and less than the USA's ID of 4
Question: How do I select all a nation's Cities when I do not know how many there are?
My query assigns every city to every country, which is probably not what you want, but:
http://sqlfiddle.com/#!5/94d63/3
SELECT *
FROM (
SELECT
place.Title AS country_name,
place.ID AS id,
(SELECT MIN(ID)
FROM place AS next_place
WHERE next_place.ID > place.ID
AND next_place.Content IS NULL
) AS next_id
FROM place
WHERE place.Content IS NULL
) AS country
INNER JOIN place
ON place.ID > country.id
AND CASE WHEN country.next_id IS NOT NULL
THEN place.ID < country.next_id
ELSE 1 END
select * from tbl
where id > 1
and id < (select min(id) from tbl where content is null and id > 1)
EDIT
I just realized the above does not work if there are no countries with greater ID. This should fix it.
select * from tbl a
where id > 4
and id < (select coalesce(b.id,a.id+1) from tbl b where b.content is null and b.id > a.id)
Edit 2 - Also made subquery fully correlated, so only have to change country id in one place.
You have here severals things to consider, one is if your data is gonna change and the other one is if it isn't gonna change, for the first one exist 2 solutions, and for the second, just one.
If your data is organize as shown in your example, you can do a select top 3, i.e.
SELECT * FROM CITIES WHERE ID NOT IN (SELECT TOP 3 ID FROM CITIES)
You can create another table where you specify wich city belongs to what parent, and make the hierarchy by yourself.
I reccomend the second one to be used.
I just learned about COALESCE and I'm wondering if it's possible to COALESCE an entire row of data between two tables? If not, what's the best approach to the following ramblings?
For instance, I have these two tables and assuming that all columns match:
tbl_Employees
Id Name Email Etc
-----------------------------------
1 Sue ... ...
2 Rick ... ...
tbl_Customers
Id Name Email Etc
-----------------------------------
1 Bob ... ...
2 Dan ... ...
3 Mary ... ...
And a table with id's:
tbl_PeopleInCompany
Id CompanyId
-----------------
1 1
2 1
3 1
And I want to query the data in a way that gets rows from the first table with matching id's, but gets from second table if no id is found.
So the resulting query would look like:
Id Name Email Etc
-----------------------------------
1 Sue ... ...
2 Rick ... ...
3 Mary ... ...
Where Sue and Rick was taken from the first table, and Mary from the second.
SELECT Id, Name, Email, Etc FROM tbl_Employees
WHERE Id IN (SELECT ID From tbl_PeopleInID)
UNION ALL
SELECT Id, Name, Email, Etc FROM tbl_Customers
WHERE Id IN (SELECT ID From tbl_PeopleInID) AND
Id NOT IN (SELECT Id FROM tbl_Employees)
Depending on the number of rows, there are several different ways to write these queries (with JOIN and EXISTS), but try this first.
This query first selects all the people from tbl_Employees that have an Id value in your target list (the table tbl_PeopleInID). It then adds to the "bottom" of this bunch of rows the results of the second query. The second query gets all tbl_Customer rows with Ids in your target list but excluding any with Ids that appear in tbl_Employees.
The total list contains the people you want — all Ids from tbl_PeopleInID with preference given to Employees but missing records pulled from Customers.
You can also do this:
1) Outer Join the two tables on tbl_Employees.Id = tbl_Customers.Id. This will give you all the rows from tbl_Employees and leave the tbl_Customers columns null if there is no matching row.
2) Use CASE WHEN to select either the tbl_Employees column or tbl_Customers column, based on whether tbl_Customers.Id IS NULL, like this:
CASE WHEN tbl_Customers.Id IS NULL THEN tbl_Employees.Name ELSE tbl_Customers.Name END AS Name
(My syntax might not be perfect there, but the technique is sound).
This should be pretty performant. It uses a CTE to basically build a small table of Customers that have no matching Employee records, and then it simply UNIONs that result with the Employee records
;WITH FilteredCustomers (Id, Name, Email, Etc)
AS
(
SELECT Id, Name, Email, Etc
FROM tbl_Customers C
INNER JOIN tbl_PeopleInCompany PIC
ON C.Id = PIC.Id
LEFT JOIN tbl_Employees E
ON C.Id = E.Id
WHERE E.Id IS NULL
)
SELECT Id, Name, Email, Etc
FROM tbl_Employees E
INNER JOIN tbl_PeopleInCompany PIC
ON C.Id = PIC.Id
UNION
SELECT Id, Name, Email, Etc
FROM FilteredCustomers
Using the IN Operator can be rather taxing on large queries as it might have to evaluate the subquery for each record being processed.
I don't think the COALESCE function can be used for what you're thinking. COALESCE is similar to ISNULL, except it allows you to pass in multiple columns, and will return the first non-null value:
SELECT Name, Class, Color, ProductNumber,
COALESCE(Class, Color, ProductNumber) AS FirstNotNull
FROM Production.Product
This article should explain it's application:
http://msdn.microsoft.com/en-us/library/ms190349.aspx
It sounds like Larry Lustig's answer is more along the lines of what you need though.
database table like this
============================
= suburb_id | value
= 1 | 2
= 1 | 3
= 2 | 4
= 3 | 5
query is
SELECT COUNT(suburb_id) AS total, suburb_id
FROM suburbs
where suburb_id IN (1,2,3,4)
GROUP BY suburb_id
however, while I run this query, it doesn't give COUNT(suburb_id) = 0 when suburb_id = 0
because in suburbs table, there is no suburb_id 4, I want this query to return 0 for suburb_id = 4, like
============================
= total | suburb_id
= 2 | 1
= 1 | 2
= 1 | 3
= 0 | 4
A GROUP BY needs rows to work with, so if you have no rows for a certain category, you are not going to get the count. Think of the where clause as limiting down the source rows before they are grouped together. The where clause is not providing a list of categories to group by.
What you could do is write a query to select the categories (suburbs) then do the count in a subquery. (I'm not sure what MySQL's support for this is like)
Something like:
SELECT
s.suburb_id,
(select count(*) from suburb_data d where d.suburb_id = s.suburb_id) as total
FROM
suburb_table s
WHERE
s.suburb_id in (1,2,3,4)
(MSSQL, apologies)
This:
SELECT id, COUNT(suburb_id)
FROM (
SELECT 1 AS id
UNION ALL
SELECT 2 AS id
UNION ALL
SELECT 3 AS id
UNION ALL
SELECT 4 AS id
) ids
LEFT JOIN
suburbs s
ON s.suburb_id = ids.id
GROUP BY
id
or this:
SELECT id,
(
SELECT COUNT(*)
FROM suburb
WHERE suburb_id = id
)
FROM (
SELECT 1 AS id
UNION ALL
SELECT 2 AS id
UNION ALL
SELECT 3 AS id
UNION ALL
SELECT 4 AS id
) ids
This article compares performance of the two approaches:
Aggregates: subqueries vs. GROUP BY
, though it does not matter much in your case, as you are querying only 4 records.
Query:
select case
when total is null then 0
else total
end as total_with_zeroes,
suburb_id
from (SELECT COUNT(suburb_id) AS total, suburb_id
FROM suburbs
where suburb_id IN (1,2,3,4)
GROUP BY suburb_id) as dt
#geofftnz's solution works great if all conditions are simple like in this case. But I just had to solve a similar problem to generate a report where each column in the report is a different query. When you need to combine results from several select statements, then something like this might work.
You may have to programmatically create this query. Using left joins allows the query to return rows even if there are no matches to suburb_id with a given id. If your db supports it (which most do), you can use IFNULL to replace null with 0:
select IFNULL(a.count,0), IFNULL(b.count,0), IFNULL(c.count,0), IFNULL(d.count,0)
from (select count(suburb_id) as count from suburbs where id=1 group by suburb_id) a,
left join (select count(suburb_id) as count from suburbs where id=2 group by suburb_id) b on a.suburb_id=b.suburb_id
left join (select count(suburb_id) as count from suburbs where id=3 group by suburb_id) c on a.suburb_id=c.suburb_id
left join (select count(suburb_id) as count from suburbs where id=4 group by suburb_id) d on a.suburb_id=d.suburb_id;
The nice thing about this is that (if needed) each "left join" can use slightly different (possibly fairly complex) query.
Disclaimer: for large data sets, this type of query might have not perform very well (I don't write enough sql to know without investigating further), but at least it should give useful results ;-)