| Oracle Index [message #332510] |
Tue, 08 July 2008 15:20  |
prax_14
Messages: 64
Registered: July 2008
|
Member |
|
|
What makes the difference between creating Index in the below two example.
Example 1 :
CREATE INDEX my_index ON table1(column1,column2)
/
Example 2:
CREATE INDEX my_index1 ON table1(column1)
/
CREATE INDEX my_index2 ON table1(column2)
/
Thanks In advance
|
|
|
|
| Re: Oracle Index [message #332512 is a reply to message #332510] |
Tue, 08 July 2008 15:27  |
Bill B
Messages: 1971
Registered: December 2004
|
Senior Member |
|
|
|
The first instance would work if you specified column1, but would not be used if you specified only column2 in your where clause. The second set of indexes would work for either, but would be slower if you specified column1 and column2 because an index merge might need to be done.
|
|
|
|
|
|
|
|
| Re: Oracle Index [message #332517 is a reply to message #332515] |
Tue, 08 July 2008 15:42 |
ehegagoka
Messages: 493
Registered: July 2005
|
Senior Member |
|
|
i was going to ask this also, but i'm just gonna add also this question. is it ok to do this
CREATE INDEX my_index ON table1(column1,column2)
/
then
CREATE INDEX my_index2 ON table1(column2)
/
just in case you just want to use the column2 as the filter?
|
|
|
|
| Re: Oracle Index [message #332547 is a reply to message #332517] |
Tue, 08 July 2008 23:15 |
Frank
Messages: 7901
Registered: March 2000
|
Senior Member |
|
|
Because indexes are not "for free" (eg. inerts are slowed down), one generally does not create indexes 'just in case'.
If you know upfront that you will use column2 without column1 in where-clauses, then it's ok.
An alternative could be to change the other index to column2, column1. But of course that could only be done if you don't select on column1 alone.
|
|
|
|
] 
Blog
Search
Help
Members
Register
Login
Home
