On Jul 12, 2008, at 1:27 PM, Antony T Curtis wrote:
On 12 Jul 2008, at 07:15, Brian Aker wrote:
Hi!
On Jul 12, 2008, at 3:03 AM, Antony T Curtis wrote:
Remove support for 64 bit unsigned type... It would remove a bunch
of special-case handling..
If you want to support 64bit unsigned, consider adding in support
for 128bit signed type (HUGEINT?).
128bit number could just be a UUID. This came up as a topic last
night at a party I was at. There has been some talk about
restructuring the class a bit, and I believe when I get around to
that, I will go on and add this type. I also want to benchmark our
UUID() generator vs the one shipped with Linux/OSX. From what I am
hearing it is much slower then the daemon based one that comes on
the system.
UUID generation must be made good. It would be very useful for
people in multi-master situation where they can have a column
declared as:
my_id HUGEINT NOT NULL PRIMARY KEY AUTO_UUID,
Where AUTO_UUID is an alternative to AUTO_INCREMENT and the UUID
would be returned in LAST_INDEX_ID() as expected. Then no need to
worry about sharding the primary key: It should be very unlikely for
pk collision.
I like that, even tho it adds to the parser.
I've thought since the very beginning of my exposure to MySQL that it
should have something like this.
Brian pointed out last night, if we generate UUIDs slightly different
from the spec (the spec says "random number plus MAC") and generate
UUIDs as MAC plus random number, then engines that do primary key
prefix compression will be able to compress the UUIDs.
--
Mark Atwood <http://mark.atwood.name>
_______________________________________________
Mailing list: https://launchpad.net/~drizzle-discuss
Post to : [email protected]
Unsubscribe : https://launchpad.net/~drizzle-discuss
More help : https://help.launchpad.net/ListHelp
