Reputation: 101
I have this 2 query :
Query #1
SELECT A1.clrn_id ,A1.gpgroup ,A1.cl_id
FROM p_dtl A1
WHERE A1.PYMT_DT = TO_DATE(:1 ,'YYYY-MM-DD')
and A1.clrn_id in (
select clrn_id
from gp_cl_rn run
where run.clrn_id = a1.clrn_id
and run.finalized_ind = :2
)
AND cl_id IN
(
SELECT cl_id
FROM hm_sal
WHERE oprt_id ='004038'
AND runctrl_id = :3
)
Query #2
SELECT B.eid
FROM JBB B
WHERE B.eid IN
(
(
SELECT eid
FROM pbank A
WHERE (bankid,branchid) IN
(
SELECT bankid ,branchid
FROM pbranch
WHERE sourceid = :1
AND estat=:2
)
AND ESTAT = :2
AND acct_type = :3
AND acct_id = (
select max(D.acct_id)
from pbank D
where D.eid = A.eid
AND D.ESTAT = :2
AND D.acct_type = :3)
)
)
How can I change the clrn_id
for the 1st query and B.eid
for the 2nd query into EXISTS function? And are the bankid
, branchid
and acct_id
also changeable into EXISTS function?
Thanks in advance!
Upvotes: 2
Views: 398
Reputation: 52843
You can change this to a where exists
, but you probably want to join
instead:
SELECT B.eid
FROM JBB B
JOIN ( SELECT eid, estat, acct_type, acct_id
, max(acct_id) over ( partition by eid ) as max_acct
FROM pbank
) A
ON b.eid = a.eid
JOIN pbranch C
ON a.bankid = c.bankid
AND a.branchid = c.brahchid
WHERE c.sourceid = :1
AND c.estat = :2
AND a.ESTAT = :3
AND a.acct_type = :4
AND a.acct_id = a.max_acct
By using the analytic function max()
you remove the need for a sub-select; it's also a lot clearer, I think, to do things in this manner.
And, your newly added query becomes:
SELECT A1.clrn_id, A1.gpgroup, A1.cl_id
FROM p_dtl A1
JOIN gp_cl_rn run
ON A1.clrn_id = run.clrn_id
WHERE A1.PYMT_DT = TO_DATE(:1 ,'YYYY-MM-DD')
AND run.clrn_id = a1.clrn_id
AND run.finalized_ind = :2
I notice that you're explicitly using the to_date()
function, which implies that you're storing a date as a string. This is bad practice and likely to cause you trouble in the longer run... avoid it if at all possible.
max(acct_id) over ( partition by eid )
is an analytic function; this does exactly the same as the aggregate function max()
, except instead of requiring a GROUP BY, it returns the same result for every row in the result set.
This particular use returns the maximum acct_id
for every eid
. There's a whole host of analytic functions, but the best thing to do is to try it for yourself. There are also several examples available online.
Using a JOIN is not necessarily quicker than a where exists
, it just depends what you're after. As with everything I would recommend trying both and seeing what suits your particular situation more.
Generally, they have different purposes; where exists
is designed to stop "calculating" rows when a single row that fulfils the conditions is found. JOIN, does everything. In your case as you want everything there may be little to chose between them but I would always use JOIN; just ensure that your indexes are correct.
Upvotes: 3
Reputation: 25753
I hope, it will help you:
SELECT B.eid
FROM JBB B
where exists
(
SELECT 1
FROM pbank A
WHERE exists
(
SELECT 1
FROM pbranch PB
WHERE PB.sourceid = :1
AND PB.estat=:2
and PB.bankid = A.bankid
AND PB.branchid = A.branchid
)
AND ESTAT = :2
AND acct_type = :3
AND A.eid = B.eid
AND acct_id = (
select max(D.acct_id)
from pbank D
where D.eid = A.eid
AND D.eid = B.eid
AND D.ESTAT = :2
AND D.acct_type = :3)
)
Upvotes: 2
Reputation: 6401
SELECT B.eid
FROM JBB B
JOIN
(
(
SELECT eid
FROM pbank A
WHERE (bankid,branchid) IN
(
SELECT bankid ,branchid
FROM pbranch
WHERE sourceid = 'BNIATB'
AND estat='A'
)
AND ESTAT = 'A'
AND acct_type = 'S'
AND acct_id = (
select max(D.acct_id)
from pbank D
where D.eid = A.eid
AND D.ESTAT = 'A'
AND D.acct_type = 'S')
)
) C on B.eid = c.EID
Upvotes: 1