El Miércoles, 11 de Febrero de 2009, Frederik Ramm escribió: > Hi, > > Iván Sánchez Ortega wrote: > > I just finished coding a bulk importer which interfaces directly to the > > MySQL DB instead of interfacing to the API. > > Well then I say let's just politely ask TomH to open up the MySQL TCP > port for public access and create user accounts for us all, and off we > go with factor 13 speed. I'm sure everyone will be pleased ;-)
You tell the french guys and their couple million buildings ;-) > Jokes aside, does your bulk importer also implement populating the > history tables, making sure that every node referenced by a way (and > every node/way/relation referenced by a relation) actually exists, and > all these totally useless integrity constraints? I'm importing brand new data, so populating the history tables is no big deal. On second thought, I haven't implemented integrity checks yet - but they would be only neccesary if a new way would reference an existing node, not a placeholder ID. (Yes, I'm trusting the optimistic locking so I'm supposing that the data that I just commited to the DB didn't get deleted so my placeholder IDs don't get screwed around - or I could just prepare a humongous transaction to make sure the diff upload is kind of atomic) <joke> Anyway, I'm a MySQL guy, not a PGSQL guy. We don't need no stinkin' integrity checks nor foreign keys here. </joke> If I have time to code the next few months, this is gonna get fun... :-) -- ---------------------------------- Iván Sánchez Ortega <[email protected]> Justifica tus limitaciones, y ciertamente las tendrás.- Richard Basch.
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ dev mailing list [email protected] http://lists.openstreetmap.org/listinfo/dev
