As you can read in README, osmo itself does not ensure any additional index (next to the default _id
which has no meaning), you must do that. If you want to, you can set p, id, v as index. It's up to
you to set the index you need. osmo itself does net check the data, so there can be multiple
versions of p, id with several v. You can insert a full history dump. (osmo does not cache
anything.) I would be interested how a Mongo cluster performs when holding the whole planet.
I am going to deal with indexing in osmo when .osc will be supported.
osmo is written in C++ for optimal performance. I once wrote the same code in node.js, and it is ~
20 times slower. I bet that Python is similarly slow.
Am 14.09.11 14:48, schrieb Jeffrey Ollie:
On Sat, Sep 10, 2011 at 6:05 PM, Andreas Kalsch<[email protected]> wrote:
Probably this is useful for you, too, so let me announce osmo, a performant
way to populate MongoDB with OSM data: https://github.com/akidee/osmo
Please give me some feedback, or you even want to contribute.
I've been working on something similar, except in Python. One thing
that I thought about as I was looking at the document structure you
were using would be to include the version in the primary key. That
would allow you to have the full history of an object in the database
if you wanted. It would complicate queries though so perhaps it would
best be left as an optional feature.
_______________________________________________
dev mailing list
[email protected]
http://lists.openstreetmap.org/listinfo/dev
