Join query returns duplicate rows - mysql

Question

purchase_request_master
prm_voucher_no| project_id| status_id| request_date
17 46 3 11-6-2016 0:00
18 46 3 20-6-2016 0:00
19 46 3 216-2016 0:00
purchase_request_details
prm_voucher_no| item_id| request_quantity
17 80 50
17 81 100
18 80 75
19 83 10
19 81 35
19 82 120
purchase_order_master
pom_voucher_no| prm_request_id |supplier_id
16 17 14
17 18 14
18 19 15
purchase_order_details
pom_voucher_no| approved_quantity| rate
16 50 1000
16 100 1500
17 75 150
18 10 2500
18 35 3000
18 120 1700
when I run the below query it gives 14 rows(duplicate row returning).expected out put row is 6.. Please refer below output tables..
select prm.prm_voucher_no,prm.project_id,prm.status_id,prd.requested_quantity,prd.item_id,pom.pom_voucher_no,pom.supplier_id,pod.rate,pod.approved_quantity
from purchase_request_master prm
left join purchase_request_details prd on prd.prm_voucher_no=prm.prm_voucher_no
left join purchase_order_master pom on prm.prm_voucher_no=pom.request_id
left join purchase_order_details pod on pom.pom_voucher_no=pod.pom_voucher_no
where prm.project_id=46 and ( EXTRACT(MONTH FROM prm.request_Date)=6) and (EXTRACT(YEAR FROM prm.request_Date)=2016)
group by prm.voucher_no,prm.project_id,prm.status_id,prd.requested_quantity,prd.item_id,pom.voucher_no,pom.supplier_id,pod.rate,pod.approved_quantity
order by prm.voucher_no
i tried inner join,distinct,distinct least,group by,temporary table,with clause all these method.. but no use every this gives duplicate row
How to solve this problem..
OUTPUT
prm_voucher_no| project_id| status_id|item_id|request_quantity |pom_voucher_no| supplier_id|approved_quantity | rate
17 46 3 80 50 16 14 100 1000
17 46 3 81 100 16 14 75 1500
17 46 3 80 75 16 15 10 150
17 46 3 81 10 16 14 35 10
18 46 3 81 35 17 14 120 35
19 46 3 80 120 18 15 50 120
19 46 3 81 50 18 14 100 1000
19 46 3 82 100 18 14 75 1500
19 46 3 80 75 18 15 10 150
19 46 3 81 10 18 14 35 10
19 46 3 82 35 18 14 120 35
19 46 3 80 120 18 15 35 120
19 46 3 81 35 18 14 50 1500
19 46 3 82 50 18 15 100 1700
EXPECTED OUTPUT
prm_voucher_no| project_id| status_id| item_id| request_quantity| pom_voucher_no| supplier_id|approved_quantity| rate
17 46 3 80 50 16 14 100 1000
17 46 3 81 100 16 14 75 1500
18 46 3 81 35 17 14 120 35
19 46 3 80 120 18 15 50 120
19 46 3 81 50 18 14 100 1000
19 46 3 82 100 18 14 75 1500

Answer 1

I think the problem is in your data model itself. Ideally, you would have a line_number field in both of your "detail" tables, and this would be used in the join:
create table purchase_request_details (
prm_voucher_no integer,
prm_voucher_line integer, // Add this
item_id integer,
request_quantity
)
create table purchase_order_details (
pom_voucher_no integer,
pom_voucher_line integer, // and this
approved_quantity integer,
rate integer
)
And then this query would give you the results you seek:
select
prm.prm_voucher_no,prm.project_id,prm.status_id,prd.request_quantity,
prd.item_id,pom.pom_voucher_no,pom.supplier_id,pod.rate,pod.approved_quantity
from
purchase_request_master prm
left join purchase_request_details prd on
prd.prm_voucher_no=prm.prm_voucher_no
left join purchase_order_master pom on
prm.prm_voucher_no=pom.prm_request_id
left join purchase_order_details pod on
pom.pom_voucher_no=pod.pom_voucher_no and
prd.prm_voucher_line = pod.pom_voucher_line // This is the key
where
prm.project_id=46 and
EXTRACT(MONTH FROM prm.request_Date) = 6 and
EXTRACT(YEAR FROM prm.request_Date) = 2016
order by prm.prm_voucher_no
If you have no ability to control the data model, then I think the best you can do is artificially add a line number. I don't recommend this at all, as you are presupposing a lot of things, most notably that the order of records in the one table automatically correlates to the order of records in the other -- and I'm betting that's far from a guarantee.
Adding a line number would be done using the row_number() analytic, and PostgreSQL has that but MySQL does not... you have both tags in your question. Which DBMS are you using?
If you can't add line numbers, can you add item_id to your purchase_order_details table? This would likely handle your issue, unless you can have the same item on multiple lines within a purchase request/order.
In the data you have above, a join on the requested quantity (prd.request_quantity = pod.approved_quantity) fixes your issue, but I am highly confident that this would burn you when you started running it against real data.