不是 GROUP BY 表达式 - OVER(PARTITION BY)

Question

I am creating a query where it shows me the total per function per month but I am getting an error - not a GROUP BY expression.

I have this in the select statement -

sum(a.sched_hours) OVER(PARTITION BY d.function) AS total_hours

Anything I can do to show the correct results?

Answer 1

Do not use an analytic function, use the normal aggregation function instead:

SELECT function_name,
       TRUNC(dt, 'MM') AS month,
       SUM(sched_hours) AS total_hours
FROM   table_name
GROUP BY
       function_name,
       TRUNC(dt, 'MM')

Which, for the sample data:

CREATE TABLE table_name (function_name, dt, sched_hours) AS
SELECT 'A', DATE '2022-01-01', 10 FROM DUAL UNION ALL
SELECT 'A', DATE '2022-01-02', 10 FROM DUAL UNION ALL
SELECT 'A', DATE '2022-02-01', 20 FROM DUAL UNION ALL
SELECT 'A', DATE '2022-02-02', 20 FROM DUAL UNION ALL
SELECT 'B', DATE '2022-01-01', 20 FROM DUAL UNION ALL
SELECT 'B', DATE '2022-01-02', 20 FROM DUAL UNION ALL
SELECT 'B', DATE '2022-02-01', 30 FROM DUAL UNION ALL
SELECT 'B', DATE '2022-02-02', 30 FROM DUAL;

Outputs:

FUNCTION_NAME	MONTH	TOTAL_HOURS
A	01-JAN-22	20
A	01-FEB-22	40
B	01-JAN-22	40
B	01-FEB-22	60

---

If you want a grant total for each function across all the months then nest the aggregation function inside the analytic function:

SELECT function_name,
       TRUNC(dt, 'MM') AS month,
       SUM(sched_hours) AS hours,
       SUM(SUM(sched_hours)) OVER (PARTITION BY function_name) AS total_hours
FROM   table_name
GROUP BY
       function_name,
       TRUNC(dt, 'MM')

Which outputs:

FUNCTION_NAME	MONTH	HOURS	TOTAL_HOURS
A	01-JAN-22	20	60
A	01-FEB-22	40	60
B	01-JAN-22	40	100
B	01-FEB-22	60	100

db<>fiddle here