a
/ \
/ \
b c
/ \ / \
d e f g
Almost all queries for that index will be able to use the buffer for 'a' and 'b'. It's not until they get things very narrowed down that you have to actually hit the index files for specific leaves. Creating ten small tables creates ten duplicate 'a' and 'b' sections which polutes the key buffer.
-Eric
Dathan Pattishall wrote:
Use 10 smaller tables and perform a union. It's faster to look up in smaller tables then larger ones-generally. Additionally more of the key_buffer can be used for tables with the most hits over the tables with less hits, making the lookup sustain speed over time.
-----Original Message-----
From: Dan Salzer [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 11, 2005 11:46 AM
To: mysql@lists.mysql.com
Subject: Union Performance Question
Hi everyone,
I have a question regarding the performance of UNION queries:
I need to do a full-text search against a large number of rows. Is it faster to have one table with 10,000,000 text rows and perform one full-text search. Or, am I better off having 10 smaller more managable tables and performing UNION'ed full-text searches against them.
A merge table seems like it would be the ideal solution, but the limitation on full-text indexing prohibits it.
Thanks!
-Dan
--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
-- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]