Eliminating NVL in WHERE Clause

Question

Hi,
I've a condition in WHERE Clause as:
WHERE NVL(tab1.col1,'X') = NVL(tab2.col2,'X')
Now, I want to eliminate NVL to optimize the query.
Can any one suggest me, how do I do this?
Thanks,
Santhosh

Answer 1

SQL> create table test (id number,id2 number);
Table created.
SQL> create table test2 (id number,id2 number);
Table created.
SQL> create index test_indx on test(id2);
Index created.
SQL> create index test2_indx on test2(id2);
Index created.
SQL> insert into test
  2  select level,level
  3  from dual
  4  connect by level <= 10000;
10000 rows created.
SQL> update test set id2 = null where mod(id,1000) = 0;
10 rows updated.
SQL> insert into test2
  2  select level,8000+level
  3  from dual
  4  connect by level <= 10000;
10000 rows created.
SQL> update test2 set id2 = null where mod(id,1000) = 0;
10 rows updated.
SQL> explain plan for
  2  select *
  3  from test t1,test2 t2
  4  where t1.id = t2.id
  5  and nvl(t1.id2,-1) = nvl(t2.id2,-1);
Explained.
SQL> select * from table(dbms_xplan.display);
PLAN_TABLE_OUTPUT
Plan hash value: 2964437775
| Id  | Operation                   | Name  | Rows  | Bytes | Cost (%CPU)| Time     |
|   0 | SELECT STATEMENT            |       |     1 |    52 |     8   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS BY INDEX ROWID| TEST  |     1 |    26 |     2   (0)| 00:00:01 |
|   2 |   NESTED LOOPS              |       |     1 |    52 |     8   (0)| 00:00:01 |
|   3 |    TABLE ACCESS FULL        | TEST2 |     1 |    26 |     6   (0)| 00:00:01 |
|*  4 |    INDEX RANGE SCAN         | INDX1 |     1 |       |     1   (0)| 00:00:01 |
PLAN_TABLE_OUTPUT
Predicate Information (identified by operation id):
   1 - filter(NVL("T1"."ID2",(-1))=NVL("T2"."ID2",(-1)))
   4 - access("T1"."ID"="T2"."ID")
Note
   - dynamic sampling used for this statement
21 rows selected.
SQL> create index indx11 on test(nvl(id2,-1));
Index created.
SQL> create index indx12 on test2(nvl(id2,-1));
Index created.
SQL> ed
Wrote file afiedt.buf
  1  explain plan for
  2  select *
  3  from test t1,test2 t2
  4  where t1.id = t2.id
  5  and nvl(t2.id2,-1) = nvl(t1.id2,-1)
  6* and t1.id > 9000
SQL> /
Explained.
SQL> select * from table(dbms_xplan.display);
PLAN_TABLE_OUTPUT
Plan hash value: 3509839790
| Id  | Operation                     | Name  | Rows  | Bytes | Cost (%CPU)| Time     |
|   0 | SELECT STATEMENT              |       |     1 |    52 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS BY INDEX ROWID  | TEST  |     1 |    26 |     2   (0)| 00:00:01 |
|   2 |   NESTED LOOPS                |       |     1 |    52 |     3   (0)| 00:00:01 |
|   3 |    TABLE ACCESS BY INDEX ROWID| TEST2 |     1 |    26 |     1   (0)| 00:00:01 |
|*  4 |     INDEX RANGE SCAN          | INDX2 |     1 |       |     1   (0)| 00:00:01 |
|*  5 |    INDEX RANGE SCAN           | INDX1 |    50 |       |     1   (0)| 00:00:01 |
PLAN_TABLE_OUTPUT
Predicate Information (identified by operation id):
   1 - filter(NVL("T2"."ID2",(-1))=NVL("T1"."ID2",(-1)))
   4 - access("T2"."ID">9000)
   5 - access("T1"."ID"="T2"."ID")
       filter("T1"."ID">9000)
Note
PLAN_TABLE_OUTPUT
   - dynamic sampling used for this statement
24 rows selected.
user652588 wrote:Hi Blushadow,
Thanks for the reply.
You suggested me to use function based index. I'm unaware of it.
Could you please let me know, what are those and how can I use those?
And also, could you please let me know the functionality COALESCE?
Thanks,
Santhoshhttp://www.oracle.com/pls/db10g/search?remark=quick_search&word=function+based+index
http://download.oracle.com/docs/cd/B14117_01/server.101/b10759/functions019.htm#SQLRF00617