MySQL Forums
Forum List  »  Merge Storage Engine

Some big databases and merge
Posted by: Gabriel Klein
Date: June 07, 2006 07:04AM

Hello,

I'm using the merge database because we have big tables.
(Actually 31'782'055 entries... / 3.9GO)

My question is: How is the optimisation done?

I have logs that have a "day granularity" and I have an index on the day field.
These logs are "merged" by month and I make a merge on these differents myisam tables.

view_full_200511
view_full_200512
view_full_200601
view_full_200602
view_full_200603
view_full_200604
view_full_200605
view_full_now

view_full is a merged view on view_full_*

Is it optimized to make something like select count(*) from view_full_* where date<20060415 and date>=20060315;

Is the time needed to parse all table optimized (as I have an index), or might it be a serious performance problem? Is it relevant to make a temporary view on the table we are interested in and perform the query on this view (Here view_full_merge = view_full_200603, view_full_200604)?

Do you see any valid reason to make a view "per days" then merge these ~500 tables, or is it not a good idea (Most of the queries are made on a week period.)?

Thank you for your answer!

Gab

Options: ReplyQuote


Subject
Views
Written By
Posted
Some big databases and merge
5747
June 07, 2006 07:04AM


Sorry, you can't reply to this topic. It has been closed.

Content reproduced on this site is the property of the respective copyright holders. It is not reviewed in advance by Oracle and does not necessarily represent the opinion of Oracle or any other party.