Hi Martin and thank you for this useful answer. I would really prefer the second approach and its far lighter process regarding the overpass server. As I don't need to extract tags on referenced objects, it will be ok for me.
All the best, François François Lacombe fl dot infosreseaux At gmail dot com www.infos-reseaux.com @InfosReseaux 2015-09-28 9:46 GMT+02:00 Martin Raifer <[email protected]>: > There are several ways to accomplish a kind of separation between the > "matched" content and the "referenced" (geometry containing) objects. > The most basic one would be to simply divide the query into two: First > only outputting the matched content, and second outputting referenced > objects. This has the downside of putting more load on the Overpass > server than what is actually needed. A better solution would be to > output the referenced content without any tags (by using `out skel;` > after the recurse-down – see overpass-turbo's wizard how this can be > done): Then you can distinguish between the main "information objects" > which have tags and the "geometry-providing" referenced stuff which > doesn't have any tags on them. This approach also reduces network > traffic, as fewer tags means less bytes to be transferred. > > Martin > > On Sun, Sep 27, 2015 at 7:02 PM, François Lacombe > <[email protected]> wrote: >> Hi all, >> >> I'm currently trying to import some features from OSM a in a DB by >> using several Overpass queries. >> >> As for loading any referenced content (nodes in ways, members in >> relations), those requests contain "(._;>;);" statement. >> >> I'm not able so far to distinguish the request matching content and >> the referenced content since all is coming between <osm> markups. >> >> My logic dealing with requests results is also trying to convert >> referenced nodes and members the same way as the request matching >> content since the logic doesn't know which is request matching or not. >> >> Should I improve my logic or get a cool way to obtain the output below ? >> In this example, only the way is matching request statements, nodes >> are part of the way. >> <osm> >> <way id="xxxxx" ... > >> <nd ref="yyyyy"/> >> <nd ref="zzzzz"/> >> </way> >> >> <reference> >> <node id="yyyyy"/> >> <node id="zzzzz"/> >> </reference> >> </osm> >> >> The processing logic would only deal with the way and ignore the nodes. >> >> Many thanks in advance for any help, all the best, >> >> >> François Lacombe >> >> fl dot infosreseaux At gmail dot com >> www.infos-reseaux.com >> @InfosReseaux >> >> _______________________________________________ >> dev mailing list >> [email protected] >> https://lists.openstreetmap.org/listinfo/dev > > _______________________________________________ > dev mailing list > [email protected] > https://lists.openstreetmap.org/listinfo/dev _______________________________________________ dev mailing list [email protected] https://lists.openstreetmap.org/listinfo/dev
