Hi!
Has anyone ever benchmarked a CSV table against load data infile?
Cheers,
--Brian
On Nov 28, 2008, at 7:33 PM, "Sheeri K. Cabral" <[EMAIL PROTECTED]>
wrote:
On 11/28/08, Brian Aker <[EMAIL PROTECTED]> wrote:
Hi!
On Nov 28, 2008, at 3:50 PM, Sheeri K. Cabral wrote:
mysqlimport should be taken out and shot, and then shot again,
fwiw. Then shot a few more times to make sure it's dead. LOAD DATA
INFILE (with LOCAL or without) is tons more useful. But having
something in SQL format has, for me, been the most useful, because
the importing is self-contained -- I can use that file to import,
period.
Ever used mysqlimport with CSV files and multiple threads? For that
I find it very useful.
I've never really found a need to do that kind of voodoo....In most
scenarios, the expectation is that data load = not fast, and the
only times I've had it be unacceptably so, something was Really
Wrong. (like the time ZFS was using 128k pages and InnoDB was using
16k pages.....or the many times a RAID disk was dead).
My experience is not, of course, gospel. And it's very possible
that I could be doing things faster, and just am not.....
But...if it's just CSV files, can't you make them CSV tables? (if
they're partial data for an existing table, you can do a CSV file
and then INSERT INTO .... SELECT ... FROM ....)
These days most of the importing I'm doing is the dump and load of
InnoDB tables to convert to innodb_file_per_table. Good to know
about that....before Jim's comment I hadn't known you could do multi-
threaded load data infile. (one of the curses of having used it for
years is that I tend not to keep up to date on things I don't
use.....I was very surprised to learn that mysqlbinlog had features
to specify a start and stop position, I'd been grepping and using
head and tail!)
-Sheeri
_______________________________________________
Mailing list: https://launchpad.net/~drizzle-discuss
Post to : [email protected]
Unsubscribe : https://launchpad.net/~drizzle-discuss
More help : https://help.launchpad.net/ListHelp
