I have a PBF file from the complete planet and I want to insert that data into a PostGIS database. However, I would like to exclude a certain region, e.g. all data that belongs to the islands of Japan.
How would I go about this?
Removing a certain region from a pbf file
openstreetmap
Related Solutions
Magnus,
What are you intending to do with the OSM data once in a postgis table ? Depending on what you intend on doing with the data (Are you looking to do spatial analysis of highways ? Make a map of a complete city including restaurants and other POIs (points of interests) ? ), your osm2pgsql style files will be a little different.
First, osm2pgsql's style file converts features based on key of each tag (A tag provides information for each geometric element).
For example, you'll find the line in the style file: node,way highway text linear
highway is the key that is used for features including roads, highways, motorways, alleys, etc. (then the value more descriptively specifies the type of highway)
Every value that has the highway as the key will be imported into your postgis database.
The other 3 portions of the line above, don't worry about them for now, the ones that are used in the osm2pgsql style file should be fine in most uses of osm2pgsql.
However, the default osm2pgsql style file is inclusive and if you use that, you'll likely import most features in the .OSM file into your postgis database. There may be less popular tags that are used in osm (here is a list of the most used keys - there are a few on that most used keys list that are not in the default style sheet - attribution, created_by, but those keys are used to describe the source of the data, and you likely don't want to include them).
As to finding out what keys are in your .osm file, just peek around in it by opening it in josm or qgis. One way to find out what tags are used is by opening your osm file in josm, then go to search, and type:way and it will return with a column of all the keys that are used on at least one way in that file.
The disadvantage to using the default style file is that you may be including data that you don't want or need into your database; increasing the conversion time (from osm2pgsql), and possibly lengthening the time for your queries (likely only noticeable if you're doing a very large, country-wide), and increasing the size of your database as well.
I'd recommend to just include the keys that you would need.
The new GDAL Version 1.10 supports Openstreetmap data:
http://www.gdal.org/ogr/drv_osm.html
If you are on Windows, you can get GDAL 1.10 precompiled from http://www.gisinternals.com/sdk/.
OSGeo4W does not yet have the 1.10 version.
Best Answer
You should take a look at the osmium-extract command from the "osmium tool" suite.
One of the example config files in (https://github.com/osmcode/osmium-tool/blob/master/extract-example-config/extracts.json) already include two examples on how to exclude a certain region from a larger area:
If your case, you would need to define an outer polygon ring covering the whole globe, then an inner ring with the areas you want to exclude.