Reputation: 775
Is innodb_file_per_table
faster than innodb_data_file_path
in select ?
Upvotes: 0
Views: 383
Reputation: 4158
I seem to remember from a previous similar question that any speed increase will maily be dependent upon the type of hardware you are running on. Best way to tell is to benchmark.
Upvotes: 0
Reputation: 30546
The Link provided by shakti Singh is good. I'll add one important point, an Innodb file never descrease his size. So the day you'll have an exploding table, even if you clean this table your innodb files will stay that big. With one file per innodb table the solution will be a backup table+drop table+import backup. It will be more complex with innob_data_file_path.
The real question you ask is 'faster' or not. This is difficult to answer. Reality is never simple. Just one example, say you have 8Go of available memory, 7Go of often used data and indexes and 15Go of rarely used tables. With innodb files separation you'll certainly have the often used innodb files completly stored in the recent disk access memory pages buffers (on a Linux system). And this will speed up things. You could not have this with one Innodb file containing 23Go data spread on several parts of the file...
Upvotes: 0
Reputation: 86476
Refer this url good explained
http://www.pythian.com/news/1067/
Upvotes: 1