No, they aren't. 1. records can go in any order, nodes / dense nodes blocks can mix; 2. IDs can go in any order; 3. relations can be circular in any direction.
If you want, you can ensure first two ones by prepeocessing input file with something like osmconvert. вт, 21 апр. 2015 г. в 17:29, François Battail <francois.batt...@sipibox.fr>: > Dear List, > > I'm developing an application like osm2pgsql but with a very different > schema, I would like to know if I can assume the following things when > parsing an osm.pbf file : > > 1. Records will always be in this order, that is: > - header, > - nodes, > - dense nodes, > - ways, > - relations. > > 2. OSM IDs are always in increasing order when parsing osm.pbf records > forward. > > 3. A relation with ID x referencing another relation with ID y is such > that x > y. > > Are these points safe to assume ? > > Thanks a lot, best regards. > > _______________________________________________ > dev mailing list > dev@openstreetmap.org > https://lists.openstreetmap.org/listinfo/dev >
_______________________________________________ dev mailing list dev@openstreetmap.org https://lists.openstreetmap.org/listinfo/dev