For Teradata-specific functions, GROUP BY determines the partitions over which the function executes. The clause does not collapse all rows with the same value for the group-by columns into a single row. Thus, the GROUP BY clause in these cases need only specify the partitioning column for the function.
For example, the following statement computes the running sales for each store by using the GROUP BY clause to partition the data in sales_tbl by StoreID:
SELECT StoreID, Sales, CSUM(Sales, StoreID) FROM sales_tbl GROUP BY StoreID;
The results look like this:
StoreID Sales CSum(Sales,StoreID) ------- -------- ------------------- 1001 1100.00 1100.00 1001 400.00 1500.00 1001 1000.00 2500.00 1001 2000.00 4500.00 1002 500.00 500.00 1002 1500.00 2000.00 1002 2500.00 4500.00 1003 1000.00 1000.00 1003 3000.00 4000.00