How to optimize that

Question

Dear Readers,
I've got query which runs 12s, is there any possibility to optymize
it ?
  SELECT distinct g.cl_id,
         cl_bank_id,
         cl_name,
         cl_corresp_name,
        cl_city,
         cl_corresp_city,
         cl_street,
         cl_corresp_street,
         cl_postal_code,
         cl_corresp_postal_code,
        0 as au_free_pass,
         'NO' as dv_main
    FROM ib_clients g, ib_auth_devices, ib_clients_replicated,
         (select dv_cl_id as cl_id, count(au_number) as c
          from ib_auth_cards, ib_auth_devices
         where au_dv_id = dv_id and au_card_state = 'NOTUSED'
          group by dv_cl_id) cards,
          (select dv_cl_id as cl_id, count(to_number) as c
          from ib_tokens, ib_auth_devices
          where to_dv_id = dv_id
          group by dv_cl_id) tokens
   WHERE g.cl_id = cp_cl_id AND
         g.cl_id = dv_cl_id AND
         g.cl_lock in ('ACTIVE', 'SOMET', 'LOGIN') AND
         dv_main = 'NO' and
         cards.cl_id (+)= g.cl_id and
         cards.c is null and
         tokens.cl_id (+)= g.cl_id and
         tokens.c is null
order by 1
568 wierszy zostało wybranych.
Plan wykonywania
   0      SELECT STATEMENT Optimizer=CHOOSE (Cost=6876 Card=79532 Byte
          s=15747336)
   1    0   SORT (UNIQUE) (Cost=4577 Card=79532 Bytes=15747336)
   2    1     HASH JOIN (Cost=2278 Card=79532 Bytes=15747336)
   3    2       TABLE ACCESS (FULL) OF 'IB_AUTH_DEVICES' (Cost=150 Car
          d=169682 Bytes=1866502)
   4    2       FILTER
   5    4         HASH JOIN (OUTER)
   6    5           FILTER
   7    6             HASH JOIN (OUTER)
   8    7               HASH JOIN (Cost=359 Card=39979 Bytes=5876913)
   9    8                 TABLE ACCESS (FULL) OF 'IB_CLIENTS' (Cost=54
           Card=39979 Bytes=599685)
  10    8                 TABLE ACCESS (FULL) OF 'IB_CLIENTS_REPLICATE
          D' (Cost=178 Card=94895 Bytes=12526140)
  11    7               VIEW (Cost=162 Card=2688 Bytes=53760)
  12   11                 SORT (GROUP BY) (Cost=162 Card=2688 Bytes=45
          696)
  13   12                   NESTED LOOPS (Cost=150 Card=2688 Bytes=456
          96)
  14   13                     TABLE ACCESS (FULL) OF 'IB_AUTH_DEVICES'
           (Cost=150 Card=339364 Bytes=4072368)
  15   13                     INDEX (UNIQUE SCAN) OF 'IB_PK_TO_DV_ID'
          (UNIQUE)
  16    5           VIEW (Cost=1165 Card=87879 Bytes=1757580)
  17   16             SORT (GROUP BY) (Cost=1165 Card=87879 Bytes=2109
          096)
  18   17               HASH JOIN (Cost=573 Card=125916 Bytes=3021984)
  19   18                 TABLE ACCESS (FULL) OF 'IB_AUTH_CARDS' (Cost
          =65 Card=125916 Bytes=1510992)
  20   18                 TABLE ACCESS (FULL) OF 'IB_AUTH_DEVICES' (Co
          st=150 Card=339364 Bytes=4072368)
Statystyki
          0  recursive calls
         30  db block gets
     353921  consistent gets
       4977  physical reads
          0  redo size
     125965  bytes sent via SQL*Net to client
       4533  bytes received via SQL*Net from client
         39  SQL*Net roundtrips to/from client
         21  sorts (memory)
          1  sorts (disk)
        568  rows processedRegards.
Grzegorz

Answer 1

WHERE
dv_main = 'NO' and
g.cl_lock in ('ACTIVE', 'SOMET', 'LOGIN') AND
cards.c is null and
tokens.c is null and
g.cl_id = cp_cl_id AND
g.cl_id = dv_cl_id AND
cards.cl_id (+)= g.cl_id and
tokens.cl_id (+)= g.cl_id
order by 1
Is this re-arranging will enchance the performance?
like user defind seraches will first execute or it doesn't matter??