Execution plan change when querying count(*)

Question

Hi,
Could any one please help me....
Here are my queries
1. select * from my_vew
2. select count(*) from my_view
Explain plan on the above queries are giving differnt plans. I would like to get the plan for count(*) as exactkly same as query no 1. Even I tried explain plan for the below query, which is also not giving the same plan..
select count(*) from (select * from my_view);
All I am trying to do here is to find the total run time without spending much time on passing the result to the client over a network (I beleive set autotrace traceonly option is also passing the result to the network.. only thing is that traceonly option wont display the results). count(*) query is chaning the plan in many ways such as index scan in place of full_table scan and etc.,
Thanks in advance.
Vasanth

Answer 1

why you want select count(*) from <table>,cause count(*) did single call neednt arraysize to be big which in turn less
LIO,when less LIO then high throughput which is not the case with select * from <table>,i dont know the internals of count(*)
but there would be some mechanics behind this code which always return with less IO
SQL> conn scott/tiger
Connected.
SQL> column plan_plus_exp format a100
SQL> set linesize 1000
SQL> desc t
Name                                                                                                                  
A                                                                                                                     
OBJECT_NAME                                                                                                           
SQL> select index_name from user_indexes where lower(table_name)='t'
  2  /
no rows selected
SQL> select count(*) from t
  2  /
  COUNT(*)
     23486
SQL> set autotrace traceonly
SQL> select * from t
  2  /
23486 rows selected.
Execution Plan
   0      SELECT STATEMENT Optimizer=CHOOSE (Cost=13 Card=23486 Bytes=681094)
   1    0   TABLE ACCESS (FULL) OF 'T' (Cost=13 Card=23486 Bytes=681094)
Statistics
         23  recursive calls
          0  db block gets
       1678  consistent gets
          0  physical reads
          0  redo size
     923712  bytes sent via SQL*Net to client
      17714  bytes received via SQL*Net from client
       1567  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
      23486  rows processed
SQL> select count(*) from t
  2  /
Execution Plan
   0      SELECT STATEMENT Optimizer=CHOOSE (Cost=13 Card=1)
   1    0   SORT (AGGREGATE)
   2    1     TABLE ACCESS (FULL) OF 'T' (Cost=13 Card=23486)
Statistics
          0  recursive calls
          0  db block gets
        118  consistent gets
          0  physical reads
          0  redo size
        381  bytes sent via SQL*Net to client
        499  bytes received via SQL*Net from client
          2  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
          1  rows processedyou need to compare boths query result consistent gets as well bytes sent via SQL*Net to client and bytes received via SQL*Net from client.
What yours requirment to get alls record as well as maximum throughput,its not the way to do.Just a clue try to implement bulk collect in this case.
Khurram