Paging and count of rows

Question

Hi,
I have a procedure doing paging and returning the count of rows:
As you see I used 2 select, one for paging and one for getting total row counts.
Is there a way to get rid of "SELECT count(*) into PO_TOTAL FROM TABLE1;"?
CREATE OR REPLACE PACKAGE BODY MYPACKAGE AS
TYPE T_CURSOR IS REF CURSOR;
PROCEDURE SP_PAGING
PI_STARTID IN NUMBER,
PI_ENDID IN NUMBER,
PO_TOTAL OUT NUMBER,
CUR_OUT OUT T_CURSOR
IS
BEGIN
OPEN CUR_OUT FOR
SELECT *
FROM (SELECT row_.*, ROWNUM rownum_
FROM (
SELECT column1, column2, column3 FROM TABLE1
) row_
WHERE ROWNUM <= PI_ENDID)
WHERE rownum_ >= PI_STARTID;
SELECT count(*) into PO_TOTAL FROM TABLE1;
END SP_PAGING;
END MYPACKAGE;

Answer 1

Yes, I can reproduce that:
SQL> create table emp1 as select * from emp
  2  /
Table created.
SQL> exec dbms_stats.gather_table_stats('SCOTT','EMP1');
PL/SQL procedure successfully completed.
SQL> EXPLAIN PLAN FOR
  2  SELECT  ename,
  3          job,
  4          sal,
  5          cnt
  6    FROM  (
  7           SELECT  ename,
  8                   job,
  9                   sal,
10                   count(*) over() cnt,
11                   row_number() over(order by 1) rn
12             FROM  emp1
13          )
14    WHERE rn BETWEEN 4 AND 9;
Explained.
SQL> @?\rdbms\admin\utlxpls
PLAN_TABLE_OUTPUT
Plan hash value: 1444408506
| Id  | Operation           | Name | Rows  | Bytes | Cost (%CPU)| Time     |
|   0 | SELECT STATEMENT    |      |    14 |   728 |     4  (25)| 00:00:01 |
|*  1 |  VIEW               |      |    14 |   728 |     4  (25)| 00:00:01 |
|   2 |   WINDOW BUFFER     |      |    14 |   252 |     4  (25)| 00:00:01 |
|   3 |    TABLE ACCESS FULL| EMP1 |    14 |   252 |     3   (0)| 00:00:01 |
PLAN_TABLE_OUTPUT
Predicate Information (identified by operation id):
   1 - filter("RN">=4 AND "RN"<=9)
15 rows selected.
SQL> EXPLAIN PLAN FOR
  2  SELECT *
  3  FROM (SELECT row_.*, ROWNUM rownum_
  4  FROM (
  5  SELECT ename,job,sal, count(*) over() FROM emp1
  6  ) row_
  7  WHERE ROWNUM <= 9)
  8  WHERE rownum_ >= 4;
Explained.
SQL> @?\rdbms\admin\utlxpls
PLAN_TABLE_OUTPUT
Plan hash value: 519025698
| Id  | Operation             | Name | Rows  | Bytes | Cost (%CPU)| Time     |
|   0 | SELECT STATEMENT      |      |     9 |   468 |     2   (0)| 00:00:01 |
|*  1 |  VIEW                 |      |     9 |   468 |     2   (0)| 00:00:01 |
|*  2 |   COUNT STOPKEY       |      |       |       |            |          |
|   3 |    VIEW               |      |     9 |   351 |     2   (0)| 00:00:01 |
|   4 |     WINDOW BUFFER     |      |     9 |   162 |     2   (0)| 00:00:01 |
|   5 |      TABLE ACCESS FULL| EMP1 |     9 |   162 |     2   (0)| 00:00:01 |
PLAN_TABLE_OUTPUT
Predicate Information (identified by operation id):
   1 - filter("ROWNUM_">=4)
   2 - filter(ROWNUM<=9)
18 rows selected.
SQL> However, you interpreted it incorrectly. Your query will still fetch all rows in TABLE1. Optimizer is smart enough to stop creating window buffer after 9 rows and that is what you see 9 in explain plan, while my query builds 14 row window buffer and then filters. So your query will be a bit faster.
SY.
Edited by: Solomon Yakobson on Jan 25, 2009 2:51 PM