Issue with subquery in Oracle corrected the query? [duplicate] - sql

Question

This question already has answers here:
ORA-00979 not a group by expression
(10 answers)
Closed 8 years ago.
SELECT COUNT(DISTINCT SEC.ERROR_GROUP_ID),
COUNT(DISTINCT SEC_DET.ERROR_GROUP_ID),
COUNT(DISTINCT MB.ERROR_GROUP_ID),
COUNT(DISTINCT OD.ERROR_GROUP_ID),
(SELECT COUNT (DISTINCT SEC_SCH.ERROR_GROUP_ID)
FROM SCHEMA.SECURITY SEC
LEFT OUTER JOIN SCHEMA.SECURITY_SCHEDULE SEC_SCH
ON SEC.MSD_SECURITY_ID =SEC_SCH.MSD_SECURITY_ID
WHERE SEC.MSD_SECURITY_ID IN
( SELECT DISTINCT main.MSD_SECURITY_ID
FROM SCHEMA2.Positions main
WHERE main.QUANTITY != 0
AND systimestamp >= main.eff_from_dt
AND main.eff_to_dt > systimestamp
AND systimestamp >= main.asrt_from_dt
AND main.asrt_to_dt > systimestamp
))
FROM SCHEMA.SECURITY SEC
JOIN SCHEMA.SECURITY_DETAIL SEC_DET
ON SEC.MSD_SECURITY_ID = SEC_DET.MSD_SECURITY_ID
LEFT OUTER JOIN SCHEMA.MUNI_BOND MB
ON SEC.MSD_SECURITY_ID=MB.MSD_SECURITY_ID
LEFT OUTER JOIN SCHEMA.OPTION_DETAIL OD
ON SEC.MSD_SECURITY_ID =OD.MSD_SECURITY_ID
WHERE SEC.MSD_SECURITY_ID IN
( SELECT DISTINCT main.MSD_SECURITY_ID
FROM SCHEMA2.Positions main
WHERE main.QUANTITY != 0
AND systimestamp >= main.eff_from_dt
AND main.eff_to_dt > systimestamp
AND systimestamp >= main.asrt_from_dt
AND main.asrt_to_dt > systimestamp
) ;
Error ORA-00936: missing expression
00936. 00000 - "missing expression"
*Cause:
*Action:
Error at Line: 365 Column: 3
The Nested query syntax needs to b corrected for this to work thats where i am stuck at ?.

Answer 1

This is a partial answer - I do not understand fully some of your code. I think it has issues.,
Note: ** means bold -- I messed up formatting ** is not part of this SQL.
You have to group by something. In this case:
(SELECT DISTINCT (SEC_SCH.ERROR_GROUP_ID)
FROM SCHEMA.SECURITY SEC
LEFT OUTER JOIN SCHEMA.SECURITY_SCHEDULE SEC_SCH
ON SEC.MSD_SECURITY_ID =SEC_SCH.MSD_SECURITY_ID
WHERE SEC.MSD_SECURITY_ID IN
( SELECT DISTINCT main.MSD_SECURITY_ID
FROM SCHEMA2.Positions main
WHERE main.QUANTITY != 0
AND systimestamp >= main.eff_from_dt
AND main.eff_to_dt > systimestamp
AND systimestamp >= main.asrt_from_dt
AND main.asrt_to_dt > systimestamp
)) **foo**
FROM SCHEMA.SECURITY SEC
JOIN SCHEMA.SECURITY_DETAIL SEC_DET
ON SEC.MSD_SECURITY_ID = SEC_DET.MSD_SECURITY_ID
LEFT OUTER JOIN SCHEMA.MUNI_BOND MB
ON SEC.MSD_SECURITY_ID=MB.MSD_SECURITY_ID
LEFT OUTER JOIN SCHEMA.OPTION_DETAIL OD
ON SEC.MSD_SECURITY_ID =OD.MSD_SECURITY_ID
WHERE SEC.MSD_SECURITY_ID IN
( SELECT DISTINCT main.MSD_SECURITY_ID
FROM SCHEMA2.Positions main
WHERE main.QUANTITY != 0
AND systimestamp >= main.eff_from_dt
AND main.eff_to_dt > systimestamp
AND systimestamp >= main.asrt_from_dt
AND main.asrt_to_dt > systimestamp
)
**group by foo** ;

Answer 2

Your overall query is an aggregation query without a group by, so it would be expected to return one row. You have a subquery in the select that can return multiple rows -- I suspect the problem is related to this structure.
I would suggest that you change the distinct to an aggregation function. But what? COUNT(DISTINCT SEC_SCH.ERROR_GROUP_ID)) ? MAX(SEC_SCH.ERROR_GROUP_ID))? LISTAGG(SEC_SCH.ERROR_GROUP_ID, ',') WITHIN GROUP (ORDER BY SEC_SCH.ERROR_GROUP_ID))? I don't know. It is not clear what you want for the third column.
Your entire query looks suspect. So many count(disintct) expressions often mean that you are joining along independent dimensions -- creating a cartesian product. Hard to say if that is a problem, because without sample data and desired results, your question doesn't actually say what you want to accomplish.