问题
I'm trying to count the frequency of numbers from 1 to 100 on different fields of a table.
Let's say I have the table "Results" with the following data:
LottoId Winner Second Third
--------- --------- --------- ---------
1 1 2 3
2 1 2 3
I'd like to be able to get the frequency per numbers. For that I'm using the following code:
--Creating numbers temp table
CREATE TABLE #Numbers(
Number int)
--Inserting the numbers into the temp table
declare @counter int
set @counter = 0
while @counter < 100
begin
set @counter = @counter + 1
INSERT INTO #Numbers(Number) VALUES(@counter)
end
--
SELECT #Numbers.Number, Count(Results.Winner) as Winner,Count(Results.Second) as Second, Count(Results.Third) as Third FROM #Numbers
LEFT JOIN Results ON
#Numbers.Number = Results.Winner OR #Numbers.Number = Results.Second OR #Numbers.Number = Results.Third
GROUP BY #Numbers.Number
The problem is that the counts are repeating the same values for each number. In this particular case I'm getting the following result:
Number Winner Second Third
--------- --------- --------- ---------
1 2 2 2
2 2 2 2
3 2 2 2
...
When I should get this:
Number Winner Second Third
--------- --------- --------- ---------
1 2 0 0
2 0 2 0
3 0 0 2
...
What am I missing?
回答1:
If you are using SQL Server 2005+
With
WinnerCounts As
(
Select #Numbers.Number, Count(Results.Winner) As Results
FROM #Numbers
JOIN Results
On #Numbers.Number = Results.Winner
)
, SecondCounts As
(
Select #Numbers.Number, Count(Results.Second) As Results
FROM #Numbers
JOIN Results
On #Numbers.Number = Results.Second
)
, ThirdCounts As
(
Select #Numbers.Number, Count(Results.Third) As Results
FROM #Numbers
JOIN Results
On #Numbers.Number = Results.Third
)
Select Numbers.Number, Coalesce(WinnerCounts.Results,0) As Winner, Coalesce(SecondCounts.Result,0) As Second, Coalesce(ThirdCounts.Result,0) As Third
From #Numbers
Left Join WinnerCounts
On WinnerCounts.Results = #Numbers.Number
Left Join SecondCounts
On SecondCounts.Results = #Numbers.Number
Left Join ThirdCounts
On ThirdCounts.Results = #Numbers.Number
Another possible solution which will work in older versions of SQL Server:
Select #Numbers.Number
, SUM( Case When Winners.Winner Is Not Null Then 1 Else 0 End ) As WinnerCount
, SUM( Case When Seconds.Second Is Not Null Then 1 Else 0 End ) As SecondCount
, SUM( Case When Thirds.Third Is Not Null Then 1 Else 0 End ) As ThirdCount
From #Numbers
Left Join Results As Winners
On Winners.Winner = #Numbers.Number
Left Join Results As Seconds
On Seconds.Second = #Numbers.Number
Left Join Results As Thirds
On Thirds.Third = #Numbers.Number
Group By #Numbers.Number
回答2:
You can use PIVOT and UNPIVOT.
SELECT Number, Winner, Second, Third
FROM (SELECT LottoID, Ranking, Number
FROM Lotto UNPIVOT (Number FOR Ranking
IN ([Winner], [Second], [Third])) AS unpvt) flat
PIVOT (COUNT(LottoId) FOR Ranking
IN ([Winner], [Second], [Third])) crosstab
来源：https://stackoverflow.com/questions/2641331/problem-counting-item-frequency-on-t-sql