I can't understand this code's bug
ID AccountID Quantity
1 1 10 Sum = 10
2 1 5 = 10 + 5 = 15
3 1 2 = 10 + 5 + 2 = 17
4 2 7 = 7
5 2 3 = 7 + 3 = 10
SELECT ID, AccountID, Quantity,
SUM(Quantity) OVER (PARTITION BY AccountID ) AS TopBorcT,
FROM tCariH
Seems like you expected the query to return running totals, but it must have given you the same values for both partitions of AccountID
.
To obtain running totals with SUM() OVER ()
, you need to add an ORDER BY
sub-clause after PARTITION BY …
, like this:
SUM(Quantity) OVER (PARTITION BY AccountID ORDER BY ID)
But remember, not all database systems support ORDER BY
in the OVER
clause of a window aggregate function. (For instance, SQL Server didn't support it until the latest version, SQL Server 2012.)
if you are using SQL 2012 you should try
SELECT ID,
AccountID,
Quantity,
SUM(Quantity) OVER (PARTITION BY AccountID ORDER BY AccountID rows between unbounded preceding and current row ) AS TopBorcT,
FROM tCariH
if available, better order by date column.
Query would be like this:
SELECT ID, AccountID, Quantity,
SUM(Quantity) OVER (PARTITION BY AccountID ) AS TopBorcT
FROM #Empl ORDER BY AccountID
Partition by works like group by. Here we are grouping by AccountID so sum would be corresponding to AccountID.
First first case, AccountID = 1 , then sum(quantity) = 10 + 5 + 2 =>
17 & For AccountID = 2, then sum(Quantity) = 7+3 => 10
so result would appear like attached snapshot.