Home » SQL & PL/SQL » SQL & PL/SQL » Best way to find no. of rows in a table
Best way to find no. of rows in a table [message #8820] Mon, 29 September 2003 06:08 Go to next message
rajesh
Messages: 173
Registered: November 1998
Senior Member
Hi All
Please let me know the best way to find the number of records in a table.

I know the below query. Is there any other query
select count(*) from tablename;

Thanks in Advance

Regards
Rajesh
Re: Best way to find no. of rows in a table [message #8823 is a reply to message #8820] Mon, 29 September 2003 06:22 Go to previous messageGo to next message
Smitha
Messages: 17
Registered: June 2001
Junior Member
Hi,

select count(*) from tablename -- instead of giving the count(*) give the count(column_name), the column_name must be the Key field for the Database. This improves the efficiency of the query.
Re: Best way to find no. of rows in a table [message #8825 is a reply to message #8823] Mon, 29 September 2003 06:42 Go to previous messageGo to next message
rajesh
Messages: 173
Registered: November 1998
Senior Member
Thanks Smitha ...I agree with u ... I want other way...
one more way is to create index on this table ...

i want the best way ...
Re: Best way to find no. of rows in a table [message #8835 is a reply to message #8823] Mon, 29 September 2003 11:36 Go to previous messageGo to next message
Todd Barry
Messages: 4819
Registered: August 2001
Senior Member
There is no difference at all in efficiency between count(*) and count(pk_column).
Re: Best way to find no. of rows in a table [message #8837 is a reply to message #8825] Mon, 29 September 2003 12:16 Go to previous message
Thiru
Messages: 1089
Registered: May 2002
Senior Member
I dont agree.
count(*) DOES NOT scan for all the columns in a record.All it does is to check for the existence of a row and increment its count.
count(col) gives the count of NON-NULL values of that column.Even if this column has NOT NULL constraint on it, it is NOT faster than count(*).

Infact, when there is a NOT null constraint, Oracle internally converts count(col) into count(*) for optimal execution.

--Lets see if there is any difference ..

SQL> drop table t;

Table dropped.

SQL> create table t as select * from all_objects;

Table created.

SQL> alter session set sql_trace=true;

Session altered.

SQL> select count(object_id) from t;

COUNT(OBJECT_ID)
----------------
29200

SQL> select count(*) from t;

COUNT(*)
----------
29200

SQL> alter session set sql_Trace=false;

Session altered.

********************************************************************************
-- Lets look at the TKPROF report

select count(object_id)
from
t

call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.00 0.00 0 0 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 2 0.01 0.01 0 820 0 1
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 4 0.01 0.01 0 820 0 1

Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 73

Rows Row Source Operation
------- ---------------------------------------------------
1 SORT AGGREGATE
29200 TABLE ACCESS FULL T

********************************************************************************

select count(*)
from
t

call count cpu elapsed disk query current rows
------- ------ -------- ---------- ---------- ---------- ---------- ----------
Parse 1 0.00 0.00 0 0 0 0
Execute 1 0.00 0.00 0 0 0 0
Fetch 2 0.01 0.01 0 820 0 1
------- ------ -------- ---------- ---------- ---------- ---------- ----------
total 4 0.01 0.01 0 820 0 1

Misses in library cache during parse: 1
Optimizer goal: CHOOSE
Parsing user id: 73

Rows Row Source Operation
------- ---------------------------------------------------
1 SORT AGGREGATE
29200 TABLE ACCESS FULL T

********************************************************************************

-- They are same in terms of performance.

--- Now,Lets do One more test. Lets create a unique index on this NOT NULL column and see if there's any performance difference...

SQL> create unique index t_idx on t(object_id);

Index created.

SQL> set autotrace on

SQL> analyze table t compute statistics;

Table analyzed.

SQL> select count(object_id) from t;

COUNT(OBJECT_ID)
----------------
29200

Execution Plan
----------------------------------------------------------
0 SELECT STATEMENT Optimizer=CHOOSE (Cost=20 Card=1)
1 0 SORT (AGGREGATE)
2 1 INDEX (FAST FULL SCAN) OF 'T_IDX' (UNIQUE) (Cost=20 Card
=29200)

Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
0 consistent gets
0 physical reads
0 redo size
0 bytes sent via SQL*Net to client
0 bytes received via SQL*Net from client
0 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed

SQL> select count(*) from t;

COUNT(*)
----------
29200

Execution Plan
----------------------------------------------------------
0 SELECT STATEMENT Optimizer=CHOOSE (Cost=20 Card=1)
1 0 SORT (AGGREGATE)
2 1 INDEX (FAST FULL SCAN) OF 'T_IDX' (UNIQUE) (Cost=20 Card
=29200)

Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
0 consistent gets
0 physical reads
0 redo size
0 bytes sent via SQL*Net to client
0 bytes received via SQL*Net from client
0 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed

SQL>

Both Count(non null unique col) and count(*) used FAST FULL SCAN of the Index to get the count.

I will go for count(*) if I want to get the record count as fast as possible, unless this is a static table and you have analyzed the table, in which case I can get NUM_ROWS from DBA_TABLES.

Hope this clarifies..
Thiru
Previous Topic: Joining tables where not all tables have relevant records
Next Topic: Can I call a Operating system command from a PL/SQL block
Goto Forum:
  


Current Time: Thu Apr 25 05:32:47 CDT 2024