How Can I make the following query faster

Question

Hi Guru
I want your valuable suggestion to make the following query faster.I did not write all required columns list. I gave here all those columns where I have conditon like decode,case when,or subquery
(SELECT CASE WHEN REPORTED_BY IS NULL THEN
          (SELECT INITCAP(EMP_NAME) FROM HR_EMP WHERE EMP_NO = M.EMP_NO_RADIO)
     ELSE (SELECT INITCAP(EMP_NAME) FROM HR_EMP WHERE EMP_NO = M.REPORTED_BY) END RADIOLOGIST_NAME,
     (SELECT TEAM_NAME FROM DC_TECHTEAMMST WHERE TEAM_NO = M.GROUP_NO) GROUP_NAME,
     CASE WHEN M.RESULT_ENTRY_LOCK_BY IS NOT NULL THEN 'R'
WHEN M.REPORT_DONE = 'D' THEN 'D'
WHEN M.REPORT_DONE = 'P' THEN 'P'
WHEN M.REPORT_DONE = 'F' THEN 'F'
     WHEN NVL(M.IMG_CAPTURED,'X') NOT IN ('B','Y') OR M.QA_RESULT = 'F' THEN 'S'
WHEN NVL(M.IMG_CAPTURED,'X') IN ('B','Y') AND NVL(M.QA_RESULT,'X') NOT IN ('B','P') THEN 'Q'
     wHEN NVL(M.IMG_CAPTURED,'X') IN ('B','Y') AND NVL(M.QA_RESULT,'X') IN ('B','P') THEN 'C'
END STATUS,
     (SELECT DECODE(NVL(V.DELIVERY_STATUS,'N'),'E',3,'U',2,1)
          FROM FN_VOUCHERCHD V WHERE V.VOUCHER_NO = M.VOUCHER_NO AND V.ITEM_NO = M.TEST_NO) DELIVERY_STATUS,
     trunc((start_time-order_end)*24,0)||' hr'||':'||
     decode(length(trunc(to_char(MOD((M.start_time-M.order_end)*24,1)*60),0)),2,to_char(trunc(to_char(MOD((M.start_time-M.order_end)*24,1)*60),0))
          ,1,to_char('0'||trunc(to_char(MOD((M.start_time-M.order_end)*24,1)*60),0)))||' mi' duration_order_capture,
     DECODE(R.CONFIDENTIAL_PATIENT,'Y','*',NVL(R.NAME,R.name_lang_name||' '||R.name_lang_fname)) PAT_NAME,
     FNC_PATIENTAGE(R.REG_NO,'',R.CONFIDENTIAL_PATIENT) pat_age,
     DECODE(R.CONFIDENTIAL_PATIENT,'Y','*',R.PATIENT_SEX) PAT_SEX
FROM DC_MODALITYAPPOINTMENT M,DC_TESTMST T,OP_REGISTRATION R
WHERE M.ACCESSION_NO IS NOT NULL AND NVL(M.CANCEL_FLAG,'N') = 'N'
AND (NVL(T.SP_GEN,'S') = 'S' OR NVL(M.DOC_REQ_GEN,'N') = 'Y')
AND M.TEST_NO IS NOT NULL AND M.TEST_NO = T.TEST_NO AND M.REG_NO = R.REG_NO)
How can I make the above query faster.
Query condition or indexing whatever is preferable please guide me.
The approximate data of tables
DC_MODALITYAPPOINTMENT 2,000,000
A lot of updating is going on some columns of this table.all columns are not in the select list, Insertion is happend in batch process by back-end trigger of another table.
Primary key based one column,
OP_REGISTRATION 500,000
Daily insertion on this table around 500 records,updation is not much.
Primary key based one column 'reg_no'
DC_TESTMST
Total records of this table not more than 1500.This is setup table. Insertion and updation is not much on this table also
I have to create a view based on this query .
and I have to create another view to serve another purpose.
In the 2nd view I need this query as well as I need another query by using union all operator based on a table(dc_oldresult)
which have 1,600,000 records.There is no DML on this table
SELECT      NVL((SELECT USER_DEFINE_TEST_NO FROM DC_TESTMST WHERE TEST_NO = SV_ID AND ROWNUM = 1 ),SV_ID) USER_D_EXAM_NO,
(SELECT TEST_TYPE FROM DC_TESTMST WHERE TEST_NO = SV_ID AND ROWNUM = 1 ) EXAM_TYPE,
     NVL((SELECT TEST_NAME FROM DC_TESTMST WHERE TEST_NO = SV_ID AND ROWNUM = 1),'Exam Code: '||sv_id) EXAM_NAME,
     (SELECT PAT_NAME FROM OP_REGISTRATION WHERE REG_NO = HN) PATIENT_NAME,
     (SELECT PAT_AGE FROM OP_REGISTRATION WHERE REG_NO = HN) PATIENT_AGE,
     (SELECT PAT_SEX FROM OP_REGISTRATION WHERE REG_NO = HN) PATIENT_GENDER
FROM DC_OLDRESULT
WHERE HN IS NOT NULL AND SV_ID IS NOT NULL AND UPPER(ACTIVE) = 'TRUE'
Should I make join DC_OLDRESULT, OP_REGISTRATION and DC_TESTMST? or The eixisting subquery is better?
I use OP_REGISTRATION and DC_TESTMST in both query
Thanks in advance
Mokarem

Answer 1

When your query takes too long ...