Hello.
I am trying to create a binary-format
for storing OSM-data.
(
http://wiki.openstreetmap.org/index.php/User:MarcusWolschon%5Cosmbin_draft#nodes.obm
)
I am looking for advise on how to create an on-disk index in one
dimension (element-id-offset where it is stored)
and 2 dimensions
On Fri, Oct 17, 2008 at 10:30:41AM +0200, Marcus Wolschon wrote:
I am trying to create a binary-format
for storing OSM-data.
(
http://wiki.openstreetmap.org/index.php/User:MarcusWolschon%5Cosmbin_draft#nodes.obm
)
I am looking for advise on how to create an on-disk index in one
dimension
On Fri, Oct 17, 2008 at 10:30:41AM +0200, Marcus Wolschon wrote:
I am looking for advise on how to create an on-disk index in one
dimension (element-id-offset where it is stored)
and 2 dimensions (bounding-box-nodeIDs and boundingBox-intersecting
bouding-boxes
of ways).
For the second point
On Fri, Oct 17, 2008 at 11:30:07AM +0200, Marcus Wolschon wrote:
Do you have any Idea how such an index can be constructed, given that
* the indexed dataset is mutable
* At no time can the complete index be loaded into memory (e.g. for
rebuilding the index).
Unfortunately not. I suppose once
Sascha Silbe wrote:
On Thu, Oct 16, 2008 at 09:21:54AM +1100, Brett Henderson wrote:
The biggest problem I found wasn't the actual processing of INSERT
statements, it was MySQL scaling non-linearly with the number of rows.
MyISAM tables are very fast to import regardless of number of rows,
Joachim Zobel wrote:
Am Donnerstag, den 16.10.2008, 09:21 +1100 schrieb Brett Henderson:
Do you have any numbers to show what sort of performance increase can
be expected.
I can load europe.osm into the current_* tables in less than 12 hours
(on my laptop - overpowered CPU, 5400
On Friday 17 October 2008, Sascha Silbe wrote:
On Fri, Oct 17, 2008 at 11:30:07AM +0200, Marcus Wolschon wrote:
Do you have any Idea how such an index can be constructed, given that
* the indexed dataset is mutable
See for example this follow-up paper on Hilbert R-trees:
On Fri, Oct 17, 2008 at 02:31:26PM +0200, Freek wrote:
See for example this follow-up paper on Hilbert R-trees:
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.45.9180
Interesting paper. Thanks for the pointer!
The whole point of R-trees is that they are efficient when stored in
On Friday 17 October 2008, Sascha Silbe wrote:
Actually it seems I've been limited in view by my own application. The
problem I'd have with updating data isn't how to map data updates to
tree updates (the topic covered by the paper mentioned above) but how to
efficiently do inserts and deletes
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA512
Stefan de Konink schreef:
Simone Cortesi schreef:
On Thu, Oct 16, 2008 at 12:10 AM, Stefan de Konink [EMAIL PROTECTED] wrote:
I have now a working image; what is the way to get a *working* bridge
between the VM and Hostsystem.
It is a bit
Thank you so much!
It looks better than I thought, right from the start :)
Anyway, I found and corrected some errors - so please build a new version when
you have some free time.
And please, could you tell me how to do it in the future - I mean the updates.
I don'th think it's normal to write
On Fri, Oct 17, 2008 at 03:25:59PM +0200, Freek wrote:
Generally you keep a bit of free space in all nodes to accommodate a
number of insertions and deletions, and only split or merge nodes when
they overflow or underflow so that you don't need to move around all
kinds of elements all the
On Friday 17 October 2008, Sascha Silbe wrote:
On Fri, Oct 17, 2008 at 03:25:59PM +0200, Freek wrote:
Generally you keep a bit of free space in all nodes to accommodate a
number of insertions and deletions, and only split or merge nodes when
they overflow or underflow so that you don't need
On Fri, Oct 17, 2008 at 05:20:52PM +0200, Freek wrote:
But how do you do those splits, i.e. node inserts?
Just pick a new block somewhere else where you have free space (for
example at the end of your mmap()'ed file), then split the pointers to
the child nodes evenly over the two blocks (old
On Friday 17 October 2008, Sascha Silbe wrote:
On Fri, Oct 17, 2008 at 05:20:52PM +0200, Freek wrote:
But how do you do those splits, i.e. node inserts?
Just pick a new block somewhere else where you have free space (for
example at the end of your mmap()'ed file), then split the pointers
On Fri, Oct 17, 2008 at 06:10:51PM +0200, Freek wrote:
OK, so you just use pointers (= index gets much larger)
Ah, you're using the fact that your tree is a complete fan-out-8 tree
(except for the last part)?
Exactly.
and often give up locality (even if you spread some free blocks
across
Hi all,
someone of the maintainers of osm2pgsql can explain that:
°°°
°
mapserv:/var/mapnik/scripts# ../osm2pgsql/osm2pgsql -m -c -d gis -U
mapservuser -W -s
On Friday 17 October 2008, Sascha Silbe wrote:
[constant bulk database with invalidation + dynamic database with
pointer-based indices for updates]
That might be quite a good idea in OSM practice, you just rebuild the
bulk database every so often. Perhaps you can even just use an
On Fri, 2008-10-17 at 18:55 +0200, Thomas Herzog wrote:
Hi all,
someone of the maintainers of osm2pgsql can explain that:
°°°
°
mapserv:/var/mapnik/scripts#
On Fri, Oct 17, 2008 at 07:47:14PM +0200, Freek wrote:
Looks like we have enough ideas for adding indexes and such to an OSM
binary format, so if the rest of the structure is fixed I think we can
get it off the ground (I wouldn't mind at least helping out here and
there).
OK, so what do we
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA512
Sascha Silbe schreef:
Node/way/relation name/ref/is_in:
- text files sorted by id (intended for namefinder - could be changed
into proper database entries similar to the way tags are handled)
What do you actually implement here?
On Fri, Oct 17, 2008 at 08:58:00PM +0200, Stefan de Konink wrote:
Node/way/relation name/ref/is_in:
- text files sorted by id (intended for namefinder - could be
changed
into proper database entries similar to the way tags are handled)
What do you actually implement here?
Example
On Friday 17 October 2008, Sascha Silbe wrote:
On Fri, Oct 17, 2008 at 07:47:14PM +0200, Freek wrote:
Looks like we have enough ideas for adding indexes and such to an OSM
binary format, so if the rest of the structure is fixed I think we can
get it off the ground (I wouldn't mind at least
As said. Anyone with SVN access do svn.openstreetmap.org can do this. It
only requires copying the .po file into the right directory, run the
compiler, checkin plugin and source and update wiki page to new version.
OK, thanks for the info! I don't think I need SVN access - I'll just post the
So, I've been translating JOSM in Launchpad recently. Now 63% of the lines are
translated.
Mainly the plug-ins are left, which I don't use, so I don't want to translate
something wrong.
I think it's time to make this up into a plug-in, so that I can see it in
reality, and also get some
So, I've been translating JOSM in Launchpad recently. Now 63% of the lines are
translated.
Mainly the plug-ins are left, which I don't use, so I don't want to translate
something wrong.
I think it's time to make this up into a plug-in, so that I can see it in
reality, and also get some
On Fri, 17 Oct 2008, Женя wrote:
So, I've been translating JOSM in Launchpad recently. Now 63% of the lines are
translated.
Mainly the plug-ins are left, which I don't use, so I don't want to translate
something wrong.
I think it's time to make this up into a plug-in, so that I can see it
On Fri, 17 Oct 2008, Женя wrote:
Anyway, I found and corrected some errors - so please build a new
version when you have some free time.
Anyone with SVN access can do. Better is you attach the .po file from
launchpad in the mail with your request, as this reduces the need to have
a
28 matches
Mail list logo
