Want to win a PS4? Go Premium and enter to win our High-Tech Treats giveaway. Enter to Win

x
Solved

# SQL where and index

Posted on 2013-10-28
Medium Priority
573 Views
Hi All,
What is the best way to query with
where mydate >= '20130813'
and substr(mynumber,16,1) in ('0','2','4','6','8')

table has 30million records

is there a way to index on substr(mynumber,16,1)

Thank you,
Bill
0
Question by:bcarlis
[X]
###### Welcome to Experts Exchange

Add your voice to the tech community where 5M+ people just like you are talking about what matters.

• Help others & share knowledge
• Earn cash & points
• 3
• 2
• 2
• +2

LVL 77

Assisted Solution

slightwv (䄆 Netminder) earned 1000 total points
ID: 39606099
>>is there a way to index on substr(mynumber,16,1)

Yes.  It's called a Function-Based Index (FBI):
http://docs.oracle.com/cd/E11882_01/server.112/e41573/data_acc.htm#PFGRF94785

Is mydate a varchar2 or an actual date?
0

LVL 32

Expert Comment

ID: 39606237
And is mynumber a number or a string?
0

LVL 38

Assisted Solution

Gerwin Jansen, EE MVE earned 200 total points
ID: 39606501
I'm guessing mynumber is a 16 character string, otherwise mod(mynumber,2) would be easier to find the even numbers instead of using substr :)

@bcarlis - Can you post a few rows of the table you're referring to? That would make it easier for us to suggest a solution for you. Change data if you need to.
0

LVL 32

Assisted Solution

awking00 earned 200 total points
ID: 39606564
gerwinjansen,
>>I'm guessing mynumber is a 16 character string, otherwise mod(mynumber,2) would be easier<<
I guessed that's the case as well. However, if not, the 16th digit would have to be the last in order for modulus to work in all cases. My real question should have been, "Is the 16th character always numeric?" In which case, the function-based indexed might be created as a bit-mapped index since the values would be limited to 0-9 and might perform better than a b-tree index. I also agree that posting sample data would be a big help in offering a solution.
0

LVL 35

Expert Comment

ID: 39608776
Please tell us the datatypes for the two columns you need to query on.  Also, please tell us which columns of the table already have indexes on them.
0

LVL 2

Author Comment

ID: 39615413
Actually I found out that all dates are actually numbers and the variable mynumber is actually varchar2
0

LVL 77

Expert Comment

ID: 39615420
Try the function based index that I mentioned above.

As long as the cardinality on the substr is decent it should use the index.
0

LVL 35

Accepted Solution

Mark Geerlings earned 600 total points
ID: 39615751
If your "date" columns are actually datatype: number, and they contain values in the format YYYYMMDD, and they are already indexed, you should refer to them like this:
where mydate >= 20130813

It is important that you *DO NOT* use single quotes around these literal values, since Oracle expects "number" values to be provided that way.  If you do use single quotes, Oracle will consider that to be a character value, so it will be forced to do an implicit datatype conversion on the number column values in the table, and this will prevent the use of existing indexes on your "date" columns (that are actually "number" datatype).

And yes, a function-based index using: "substr(mynumber,16,1)" should work well for that column.

Create that index something like this:

create index last_digit_ix on [table_name] (substr(mynumber,16,1))
tablespace [tablespace_name] parallel 4 nologging;

Note: you can use any name you like for that index, I used "last_digit_ix" as an example.  Provide the actual table_name in place of [table_name].  And, you don't need the last line, but you can optionally provide a tablespace_name and/or a degree of parallelism, and/or the "nologging" option (which applies to the initial index build only).
0

LVL 77

Assisted Solution

slightwv (䄆 Netminder) earned 1000 total points
ID: 39615779
>>and this will prevent the use of existing indexes on your "date" columns (that are actually "number" datatype)

I'm not sure this is true but I've been wrong before.  Maybe in certain situations in certain versions.

Given the over-simplified test case below, both queries use the index.

``````drop table tab1 purge;
create table tab1(col1 number, col2 varchar2(20));

begin
for i in 1..10000 loop
insert into tab1 values(dbms_random.value,dbms_random.string('a',20));
end loop;
end;
/

create index tab1_idx on tab1(col1);

set autotrace traceonly
select col1 from tab1 where col1 >= '1';

select col1 from tab1 where col1 >= 1;
set autortace off
``````
0

LVL 2

Author Closing Comment

ID: 39619399
Thank you for your feedback.. got the indexes in place.. much better
0

## Featured Post

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

PL/SQL can be a very powerful tool for working directly with database tables. Being able to loop will allow you to perform more complex operations, but can be a little tricky to write correctly. This article will provide examples of basic loops alon…
This month, Experts Exchange sat down with resident SQL expert, Jim Horn, for an in-depth look into the makings of a successful career in SQL.
Via a live example, show how to take different types of Oracle backups using RMAN.
This videos aims to give the viewer a basic demonstration of how a user can query current session information by using the SYS_CONTEXT function
###### Suggested Courses
Course of the Month12 days, left to enroll