Last two joins cause duplicate rows - sql

Question

Ok, so I have a query that is returning more rows than expected with repeating data. Here is my query:
SELECT AP.RECEIPTNUMBER
,AP.FOLDERRSN
,ABS(AP.PAYMENTAMOUNT)
,ABS(AP.PAYMENTAMOUNT - AP.AMOUNTAPPLIED)
,TO_CHAR(AP.PAYMENTDATE,'MM/DD/YYYY')
,F.REFERENCEFILE
,F.FOLDERTYPE
,VS.SUBDESC
,P.NAMEFIRST||' '||P.NAMELAST
,P.ORGANIZATIONNAME
,VAF.FEEDESC
,VAF.GLACCOUNTNUMBER
FROM ACCOUNTPAYMENT AP
INNER JOIN FOLDER F ON AP.FOLDERRSN = F.FOLDERRSN
INNER JOIN VALIDSUB VS ON F.SUBCODE = VS.SUBCODE
INNER JOIN FOLDERPEOPLE FP ON FP.FOLDERRSN = F.FOLDERRSN
INNER JOIN PEOPLE P ON FP.PEOPLERSN = P.PEOPLERSN
INNER JOIN ACCOUNTBILLFEE ABF ON F.FOLDERRSN = ABF.FOLDERRSN
INNER JOIN VALIDACCOUNTFEE VAF ON ABF.FEECODE = VAF.FEECODE
WHERE AP.NSFFLAG = 'Y'
AND F.FOLDERTYPE IN ('405B','405O')
Everything works fine until I add the bottom two Inner Joins. I'm basically trying to get all payments that had NSF. When I run the simple query:
SELECT *
FROM ACCOUNTPAYMENT
WHERE NSFFLAG = 'Y'
I get only 3 rows pertaining to 405B and 405O folders. So I'm only expecting 3 rows to be returned in the above query but I get 9 with information repeating in some columns. I need the exact feedesc and gl account number based on the fee code that can be found in both the Valid Account Fee and Account Bill Fee tables.
I can't post a picture of my output.
Note: when I run the query without the two bottom joins I get the expected output.
Can someone help me make my query more efficient? Thanks!
As requested, below are the results that my query is returning for vaf.feedesc and vaf.glaccountnumber columns:
Boiler Operator License Fee 2423809
Boiler Certificate of Operation without Manway - Revolving 2423813
Installers (Boiler License)/API Exam 2423807
Boiler Public Inspection/Certification (State or Insurance) 2423816
Boiler Certificate of Operation with Manway 2423801
Boiler Certificate of Operation without Manway 2423801
Boiler Certificate of Operation with Manway - Revolving 2423813
BPV Owner/User Program Fee 2423801
Installers (Boiler License)/API Exam Renewal 2423807

Answer 1

The cause is that at least one of the connections ACCOUNTBILLFEE-FOLDER or VALIDACCOUNTFEE-ACCOUNTBILLFEE is not one-to-one. It allows for one Folder to have many AccountBillFees or for one ValidAccountFee to have many AccountBillFees.
To find the cause of such a problem this is what I usually do:
Change the SELECT A, B, C part of your query to SELECT *.
Reduce the results to one of the rows that is causing you trouble (by adding a WHERE ...). That is a single row without your last two joins and a few rows after you add those two joins.
Look at the result table from left to right. The first columns will probably show the same values for all rows. Once you see a difference between the values in a column, you know that the table of the column you are currently looking at is causing your "multiple row problem".
Now create a SELECT * statement that includes only the two tables joined together that cause multiple rows with the same WHERE ... you used above.
The result should give you a clear picture of the cause.
Once you know the reason for your problem you can think of a solution ;)

Answer 2

Try this if it helps then those tables have additional rows which are not relevant. If it doesn't then look at the results of the subqueries I have below to see what additional filters are needed
SELECT AP.RECEIPTNUMBER
,AP.FOLDERRSN
,ABS(AP.PAYMENTAMOUNT)
,ABS(AP.PAYMENTAMOUNT - AP.AMOUNTAPPLIED)
,TO_CHAR(AP.PAYMENTDATE,'MM/DD/YYYY')
,F.REFERENCEFILE
,F.FOLDERTYPE
,VS.SUBDESC
,P.NAMEFIRST||' '||P.NAMELAST
,P.ORGANIZATIONNAME
,VAF.FEEDESC
,VAF.GLACCOUNTNUMBER
FROM ACCOUNTPAYMENT AP
INNER JOIN FOLDER F ON AP.FOLDERRSN = F.FOLDERRSN
INNER JOIN VALIDSUB VS ON F.SUBCODE = VS.SUBCODE
INNER JOIN FOLDERPEOPLE FP ON FP.FOLDERRSN = F.FOLDERRSN
INNER JOIN PEOPLE P ON FP.PEOPLERSN = P.PEOPLERSN
INNER JOIN
(
SELECT DISTINCT ABF.FEECODE, ABF.FOLDERRSN
FROM ACCOUNTBILLFEE ABF
) ABF ON F.FOLDERRSN = ABF.FOLDERRSN
INNER JOIN
(
SELECT DISTINCT VAF.FEEDESC, VAF.GLACCOUNTNUMBER, VAF.FEECODE
FROM VALIDACCOUNTFEE VAF
) VAF ON ABF.FEECODE = VAF.FEECODE
WHERE AP.NSFFLAG = 'Y'
AND F.FOLDERTYPE IN ('405B','405O')

Answer 3

The data for those last two tables is different in different records in the one to many relationship. Since distinct did not fix the problem, then you have to accept that 9 records is the correct return because you are returning the fields that are different or you have to determine which of the multiple records you don't want returned based on business rules that must come from someone in your company not us.
I don't think you fully understand how SQl works as 9 records is exactly what I would have expected given the information you gave in the question. The following are some queries that show how joining in a one to many relationship can affect output and ways that you can adjust the query to get rid of the duplicated output.
Note that in some of the cases, the query cannot be adjusted to get rid of the output because of the columns you want returned. So even if some of the columns are repeated, if even one of the columns you want return has differnt records and you have no approriate business rules for which of them you want to see, you can't reduce the records set. Which rules you need are based on the type of data you are querying and what the rqeuirements are. This is not a question we can answer here, only your company knows whether a min or max value would be acceptable or if you need to add a where clause and if so what field to put it on and what values to use it to exclude. Those are business rules not SQL.
create table #temp (myid int , mydescription varchar(30))
insert into #temp(myid, mydescription)
values (1, 'test') , (2, 'test2')
create table #temp2 (myid int, myotherdescription varchar(30))
insert into #temp2(myid, myotherdescription)
values (1, 'othertest') , (1, 'othertest2'), (2, 'myothertest') , (1, 'othertest3')
select *
from #temp t
join #temp2 t2 on t.myid = t2.myid
select t2.myid, t.mydescription
from #temp t
join #temp2 t2 on t.myid = t2.myid
select distinct t2.myid, t.mydescription
from #temp t
join #temp2 t2 on t.myid = t2.myid
select t.myid, t.mydescription, t2.myotherdescription
from #temp t
join #temp2 t2 on t.myid = t2.myid
select distinct t.myid, t.mydescription, t2.myotherdescription
from #temp t
join #temp2 t2 on t.myid = t2.myid
select t.myid, min(t2.myotherdescription)
from #temp t
join #temp2 t2 on t.myid = t2.myid
group by t.myid
select t.myid, t2.myotherdescription
from #temp t
join #temp2 t2 on t.myid = t2.myid
where t2.myid = 2