Can you group together multiple SQL count statements? - sql

Question

I have a database table that records transactions in a shop. The main data recorded is the cost of each item. E.g.
Item || Cost
TV || 80.00
XboxGame || 55.00
Monitor || 45.00
Controller || 15.00
I want to find out the number of items purchased that cost less than $25, how many cost between $25 and $49.99, and how many cost $50 and above. This should be the result of running the sql command:
Table1 (Target):
Cost || Number of Items
==================================
Less than $25 || 1
$25 - $49.99 || 1
$50 and above || 2
I have tried the following:
SELECT (SELECT COUNT(*) AS Expr1
FROM tblTransactions
WHERE (Cost < 25)) AS [Less than $25],
(SELECT COUNT(*) AS Expr1
FROM tblTransactions AS tblTransactions_1
WHERE (Cost >= 25) AND (Cost < 50)) AS [$25 - $49.99],
(SELECT COUNT(*) AS Expr1
FROM tblTransactions AS tblTransactolions_2
WHERE (ContributionAmount >= 50)) AS [$50 and above];
However this produces:
Table 2 (actual result)
Less than $25 || $25 - $49.99 || $50 and above
========================================
1 || 1 || 2
My main question is: how do I change my SQL so that it produces Table 1, instead of Table 2?

Answer 1

Use Group By:
Select
[Cost] = CASE WHEN cost < 25 then 'Less than $25'
WHEN cost >= 25 and cost < 50 then '$25 - $49.99'
WHEN cost >= 50 then '$50 and above' END
,COUNT(*) As [Count]
FROM tblTransactions
GROUP BY (CASE WHEN cost < 25 then 'Less than $25'
WHEN cost >= 25 and cost < 50 then '$25 - $49.99'
WHEN cost >= 50 then '$50 and above' END)
ORDER BY CASE WHEN
cost < 25 then '1'
WHEN cost >= 25 and cost < 50 then '2'
WHEN cost >= 50 then '3' END
Edited to include a sorting.

Answer 2

You need to have separate queries that are unioned together:
SELECT 'Less than $25' AS [Cost], COUNT(*) AS [Number of Items]
FROM tblTransactions
WHERE (Cost < 25)
UNION ALL
SELECT '$25 - $49.99', COUNT(*)
FROM tblTransactions AS tblTransactions_1
WHERE (Cost >= 25) AND (Cost < 50)
UNION ALL
SELECT '$50 and above', COUNT(*)
FROM Transactions AS tblCoachBudgetHistory_2
WHERE (ContributionAmount >= 50)

Answer 3

Use UNION ALL:
SELECT 'Some text' AS [Label],
COUNT(*)
FROM TABLE
WHERE SomeCol = 'some val'
UNION ALL
SELECT 'Some text' AS [Label],
COUNT(*)
FROM TABLE
WHERE SomeCol = 'some val'

Answer 4

You could do that with a UNION
(SELECT COUNT(*) AS Count,
, "Less than $25" AS Label
FROM tblTransactions
WHERE Cost < 25)
UNION ALL
(SELECT COUNT(*) AS Count
, "$25 - $49.99" as Label
FROM tblTransactions
WHERE (Cost >= 25) AND (Cost < 50))
UNION ALL
(SELECT COUNT(*) AS Count
,"$50 and above" as Label
FROM Transactions
WHERE ContributionAmount >= 50)

Answer 5

If I were writing this, I would start with a subquery (or view) to first group the values into buckets.
SELECT
(CASE WHEN (Cost >= 50) THEN 50
WHEN (Cost >= 25) THEN 25
ELSE 0 END) as costLevel
FROM transactions t
Then it's trivial to group.
SELECT
(CASE costLevel WHEN 50 THEN '$50 and above'
WHEN 25 THEN '$25 - $49.99'
ELSE 'Less than $25' END) as [Cost]
, COUNT(*) AS [Number of Items]
FROM (thePreviousQuery) AS cl
GROUP BY costLevel
-- optional:
-- ORDER BY costLevel
The above syntax is for SQL Server, so YMMV elsewhere. (I've also treated tblTransactions and Transactions as a typo.)

Answer 6

Personally, I tend to prefer range tables for this;
SELECT name, COUNT(*)
FROM (VALUES ('Less than $25', CAST(0 as DECIMAL(5, 2)), CAST(25 as DECIMAL(5, 2))),
('$25 - $49.99', 25, 50),
('$50 and above', 50, null)) Range(name, lower, upper)
LEFT JOIN Transactions
ON cost >= lower
AND (upper IS NULL OR cost < upper)
GROUP BY name
(and working fiddle example. Should work on any RDBMS, pretty much)
This has the added benefit that the table can actually be persisted, meaning maintainable by end-users. Also, an index on cost can be used by the optimizer, meaning potentially better performant queries. Due to the LEFT JOIN, you'll also get a 0 count for ranges that have no transactions.