Improving a simple select query, which uses all rows.

Question

Hi All,
Please excuse me if the question is too silly. Below is my code
SQL> select * from v$version;
BANNER
Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Prod
PL/SQL Release 10.2.0.1.0 - Production
CORE    10.2.0.1.0      Production
TNS for 32-bit Windows: Version 10.2.0.1.0 - Production
NLSRTL Version 10.2.0.1.0 - Production
Elapsed: 00:00:00.07
SQL> show parameter optim
NAME                                 TYPE        VALUE
object_cache_optimal_size            integer     102400
optimizer_dynamic_sampling           integer     2
optimizer_features_enable            string      10.2.0.1
optimizer_index_caching              integer     0
optimizer_index_cost_adj             integer     100
optimizer_mode                       string      ALL_ROWS
optimizer_secure_view_merging        boolean     TRUE
plsql_optimize_level                 integer     2
SQL> explain plan for select SUM(decode(transaction_type,'D',txn_amount,0)) payments_reversals,
  2          SUM(decode(transaction_type,'C',txn_amount,0)) payments,primary_card_no,statement_date
  3             from credit_card_pymt_dtls  group by primary_card_no,statement_date;
Explained.
SQL> select * from table(dbms_xplan.display);
PLAN_TABLE_OUTPUT
Plan hash value: 2801218574
| Id  | Operation          | Name                  | Rows  | Bytes |TempSpc| Cost (%CPU)| Time     |
|   0 | SELECT STATEMENT   |                       |  1912K|    56M|       | 21466   (3)| 00:04:18 |
|   1 |  SORT GROUP BY     |                       |  1912K|    56M|   161M| 21466   (3)| 00:04:18 |
|   2 |   TABLE ACCESS FULL| CREDIT_CARD_PYMT_DTLS |  1912K|    56M|       |  4863   (3)| 00:00:59 |
9 rows selected.
SQL> select index_name,index_type
  2  from all_indexes
  3  where table_name = 'CREDIT_CARD_PYMT_DTLS';
INDEX_NAME                     INDEX_TYPE
INDX_TRANTYPE                  BITMAP
INDX_PCARD                     NORMAL
INDX_PSTATEMENT_DATE           NORMALThe query is using all the records in the CREDIT_CARD_PYMT_DTLS table. Transaction type will be either 'C' or 'D'.
CREDIT_CARD_PYMT_DTLS has 2 million rows and the qury will output 1.5 million rows. Table statisticas are upto date.
The query now is taking almost 5 minutes. Is thaere any way to reduce the time ?
Our DB server has 8 CPUs and 8 GB memory. Is the timing genuine ?
Thanks in Advance.
Edited by: user11115924 on Apr 29, 2009 2:43 AM
All the columns used in the query are already indexed. ( Ofcourse, not only for this query.)

Answer 1

Hi All,
Thanks for the helps provided. Expecting it once more..
My actual query is as below
select     primary_card_no,base_segment_number,atab.previous_balance,current_balance,intrest_amt_due_this_cycle,total_min_amt_due,total_credit_limit,
          total_purchase_this_cycle,total_cash_trns_this_cycle,available_credit_limit,payments,utilization,payment_ratio,payments_reversals,cash_limit,
          available_cash_limit, description
from
       select primary_card_no,DECODE(base_segment_number,NULL,primary_card_no,base_segment_number) base_segment_number,
               SUM(previous_balance) previous_balance,SUM(current_balance) current_balance ,SUM(intrest_amt_due_this_cycle) intrest_amt_due_this_cycle,
               SUM(total_min_amt_due) total_min_amt_due,SUM(total_credit_limit_all) total_credit_limit,
               SUM(total_purchase_this_cycle) total_purchase_this_cycle,SUM(total_cash_trns_this_cycle) total_cash_trns_this_cycle,
               SUM(available_credit_limit) available_credit_limit,SUM(payments) payments,
               (SUM(NVL(current_balance,0)) / SUM(total_credit_limit_all)) * 100 utilization,
               (SUM(NVL(payments,0)) / DECODE(SUM(previous_balance),0,NULL,SUM(previous_balance))) * 100 payment_ratio,
               SUM(payments_reversals) payments_reversals,SUM(cash_limit) cash_limit,SUM(available_cash_limit) available_cash_limit
      from
        ( select a.*,NVL(payments_reversals,0)payments_reversals ,NVL(payments,0) payments
          from
                (  select primary_card_no,previous_balance,current_balance,intrest_amt_due_this_cycle,total_min_amt_due,total_purchase_this_cycle,
                          total_cash_trns_this_cycle,statement_date,available_credit_limit,cash_limit,available_cash_limit,
                       (case when statement_date <= TO_DATE('301108','ddmmyy') then NULLIF(total_credit_limit,0)
                                  else NULLIF((select credit_limit
                                            from ccm_dbf_chtxn_v0 t1
                                               where t1.batch_id = '011208'
                                               and SUBSTR(t1.card_number,4) = a.primary_card_no),0)
                         end) total_credit_limit_all
                 from
                    ( select primary_card_no,previous_balance,current_balance,INTREST_AMT_DUE_THIS_CYCLE,
                                TOTAL_MIN_AMT_DUE,TOTAL_PURCHASE_THIS_CYCLE,TOTAL_CASH_TRNS_THIS_CYCLE,statement_date,
                               AVAILABLE_CREDIT_LIMIT,cash_limit,available_cash_limit,total_credit_limit
                      from credit_card_master_all@FGBAPPL_LINK
                    ) a
                 where statement_date between ADD_MONTHS(TRUNC(SYSDATE,'mm'),-6) and  TRUNC(SYSDATE,'mm')-1
             ) a,
             ( select SUM(decode(transaction_type,'D',txn_amount,0)) payments_reversals,
                           SUM(decode(transaction_type,'C',txn_amount,0)) payments,primary_card_no,TO_CHAR(statement_date,'MON-RRRR') sdate
                from credit_card_pymt_dtls 
                group by primary_card_no,TO_CHAR(statement_date,'MON-RRRR')
             ) b
         where TO_CHAR(a.statement_date,'MON-RRRR')= b.sdate(+)
           and a.primary_card_no= b.primary_card_no(+)
        ) a,
        (  select SUBSTR(a.card_number,4) card_number,base_segment_number,TO_DATE(account_creation_date,'DDMMYYYY') account_creation_date,
                       a.batch_id, credit_limit credit_limit_current
            from
                   (  select *
                       from ccm_dbf_phtxn_v0
                       where batch_id=  (SELECT to_char(MAX(TO_DATE(SUBSTR(BATCH_ID,1,6),'DDMMRR')),'DDMMRR') FROM CCM_MST_V0)
                    ) a,
                    ( select *
                      from ccm_dbf_chtxn_v0
                      where batch_id=(SELECT to_char(MAX(TO_DATE(SUBSTR(BATCH_ID,1,6),'DDMMRR')),'DDMMRR') FROM CCM_MST_V0)
                    ) b
            where a.card_number=b.card_number
            and TO_NUMBER(ROUND(MONTHS_BETWEEN(SYSDATE,TO_DATE(account_creation_date,'DDMMYYYY')),2)) >=6
            and a.company ='BNK'
        ) b
   where a.primary_card_no = b.card_number
   group by primary_card_no,base_segment_number) atab, card_summary_param btab
where utilization between utilization_low and utilization_high
and payment_ratio between  payment_ratio_low and payment_ratio_high
and SIGN(atab.previous_balance) =btab.previous_balanceWhere I have to put the PARALLEL hint for maximum performance?
Sorry for asking blindly without doing any R&D. Time is not permitting that...
Edited by: user11115924 on Apr 29, 2009 5:09 AM
Sorry for the kiddy aliases.. Query is not written by me..