Forum: Ruby on Rails mySQL newbie seek help with db indexing

Announcement (2017-05-07): www.ruby-forum.com is now read-only since I unfortunately do not have the time to support and maintain the forum any more. Please see rubyonrails.org/community and ruby-lang.org/en/community for other Rails- und Ruby-related community platforms.
Jeff P. (Guest)
on 2008-12-20 08:01
Hi,
I don't quite understand the syntax for database indexes and/or how the
work to help speed up specific database searches.  Can somebody suggest
a set of indexs that would make this set of queries go faster?

  Contact Load (0.001406)   SELECT * FROM `contacts` WHERE
(`contacts`.user_id = 1)
  Phone Load (0.003268)   SELECT * FROM `phones` WHERE
(`phones`.contact_id = 8485)
  Email Load (0.000371)   SELECT * FROM `emails` WHERE
(`emails`.contact_id = 8485)
  Note Load (0.019782)   SELECT * FROM `notes` WHERE (`notes`.contact_id
= 8485)
  Group Load (0.000154)   SELECT `groups`.* FROM `groups` INNER JOIN
groupers ON groups.id = groupers.group_id WHERE ((`groupers`.contact_id
= 8486))

a Contact has_many phones, emails, notes.
a Contact has_many groups :through groupers

Thanks much, and thanks even more if you include some explanation!

thanks,
jp
Marnen L. (Guest)
on 2008-12-20 09:12
(Received via mailing list)
On Dec 20, 1:01 am, Jeff P. <removed_email_address@domain.invalid>
wrote:
> Hi,
> I don't quite understand the syntax for database indexes and/or how the
> work to help speed up specific database searches.  Can somebody suggest
> a set of indexs that would make this set of queries go faster?
[...]

Very simple: you can profit by indexing on almost anything you're
going to be doing a lot of searching on.  In this case, that means
that you should consider building indices on your foreign key fields
(such as contacts.user_id), since they're what you're searching on in
these queries.  But before you do that, you might want to try EXPLAIN
SELECT on the queries you're interested in, and see what mySQL tells
you about where the bottlenecks are.

Best,
--
Marnen Laibow-Koser
removed_email_address@domain.invalid
http://www.marnen.org
Frederick C. (Guest)
on 2008-12-20 13:33
(Received via mailing list)
On 20 Dec 2008, at 07:11, Marnen Laibow-Koser wrote:

> [...]
>
> Very simple: you can profit by indexing on almost anything you're
> going to be doing a lot of searching on.  In this case, that means
> that you should consider building indices on your foreign key fields
> (such as contacts.user_id), since they're what you're searching on in
> these queries.  But before you do that, you might want to try EXPLAIN
> SELECT on the queries you're interested in, and see what mySQL tells
> you about where the bottlenecks are.
>
All of the above is good. To explain a little more if you do

SELECT * from contacts where contacts.user_id = 2356

and there is no index then the database must examine every single row
in the contacts table to see if it has the right user_id whereas if it
has an index it can go straight there. (similar to how if you have a
phone book and you're looking for all entries 'Michael smith' you
don't have to read the entire phone book, you can jump to the
appropriate page very quickly)

Fred
Jeff P. (Guest)
on 2008-12-20 18:45
Frederick C. wrote:
> On 20 Dec 2008, at 07:11, Marnen Laibow-Koser wrote:
>
>> [...]
>>
>> Very simple: you can profit by indexing on almost anything you're
> All of the above is good. To explain a little more if you do
>
> SELECT * from contacts where contacts.user_id = 2356
>
> and there is no index then the database must examine every single row
> in the contacts table to see if it has the right user_id whereas if it
> has an index it can go straight there. (similar to how if you have a
>
> Fred

Thanks Marnen and Fred,
Does the same logic completely cover joins, or is there something
special to consider for those?

thanks,
jp
Frederick C. (Guest)
on 2008-12-20 18:47
(Received via mailing list)
On 20 Dec 2008, at 16:45, Jeff P. wrote:

>>
>> and there is no index then the database must examine every single row
>> in the contacts table to see if it has the right user_id whereas if
>> it
>> has an index it can go straight there. (similar to how if you have a
>>
>> Fred
>
> Thanks Marnen and Fred,
> Does the same logic completely cover joins
Yup

Fred
This topic is locked and can not be replied to.