Re: [OSM-dev] Can SQLite3 handle OSM 150G data file?

2009-05-15 Thread Tomas Kolda
Did you try my version of import? I have no responses, so maybe not :). I did not tried complete planet, but indexes was created quite fast. 30seconds on 150MB osm file. I think that you should make a chart (xml size-time to create index) and than you can see complexity of creating indexes in

[OSM-dev] Sqlite module for Mapnik

2009-05-04 Thread Tomas Kolda
Hi, this is reaction for threads about sqlite and Mapnik. I promissed that I post something about my research. So here it is. http://osm.w2n.cz/ There is SVN with sources and compiled versions of tools (windows only at this time). At first I must say, that this is not replacement for current

Re: [OSM-dev] Can SQLite3 handle OSM 150G data file?

2009-04-29 Thread Tomas Kolda
Hi, I will post something, but now I'm not at home. So maybe today night. But Dane wrote me about sqlite support in mapnik, so you can try that: Are you aware we have a postgres - sqlite converter in mapnik?

Re: [OSM-dev] Can SQLite3 handle OSM 150G data file?

2009-04-28 Thread Tomas Kolda
There is no reason to be sqlite slow. How much are you are using transactions on your local db? What are you using for import? I'm importing Czech republic (50Mb bziped) in one or two minutes. So for 5.4GB it takes 108 minutes (two hours). But I'm importig to modified osm schema (something

Re: [OSM-dev] Embedded databases

2009-04-28 Thread Tomas Kolda
As I wrote in another thread I tried sqlite to do rendering along with mapnik. There is no problem to do it even with builtin rtree spatial index. I can try to create some SVN repos to share code, but it was just for playing (not for production). I tweaked postgress datasource in mapnik and

[OSM-dev] Mapnik and sqlite

2008-11-12 Thread Tomas Kolda
Hi, I am working on SQLite as mapnik datasource. It will be great as zero configuration mapnik testing suite. It uses modified WKB storage as ints (not doubles) so it saves some space. It takes only about 10 percent size more than bziped OSM. It uses btree spatial index so it should be very

Re: [OSM-dev] Binary OSM; the first pass encoder

2008-11-11 Thread Tomas Kolda
Integer has better precision if you are far from zero, because floats use only 24 bits for significant, other bits is for sign and exponent. So 32 bit integer is better if you know extends (-180, 180). Min and max of int is -2.1e9 to 2.1e9. So if you multiply longitude and latitude by 1e7

Re: [OSM-dev] Binary OSM; the first pass encoder

2008-11-11 Thread Tomas Kolda
and int Tom Stefan de Konink napsal(a): On Wed, 12 Nov 2008, Tomas Kolda wrote: Integer has better precision if you are far from zero, because floats use only 24 bits for significant, other bits is for sign and exponent. So 32 bit integer is better if you know extends (-180, 180

[OSM-dev] Complete OSM history for bounding rect

2008-10-14 Thread Tomas Kolda
Hi all, is it possible to get complete history for area? For example using osmxapi, or is there some planet dump with history? I do not want to use query for each object to get history... It is slow and there is so much overhead. Thanks for help Tomas