extract for Portland, Oregon contains data outside of bounding box

[dobratzp]

I'm looking at the extract for Portland, Oregon and it looks like there is some data which doesn't belong.

portland_oregon.osm, line 19849425:

<way id="404118908" version="1" timestamp="2016-03-16T18:44:38Z" changeset="37879771" uid="2607707" user="Marcos Medeiros">
<nd ref="4063882877"/>
<tag k="name" v="Auto Elétrica Ituiutaba"/>
<tag k="shop" v="car_repair"/>
<tag k="phone" v="+55 34 3268-1246"/>
<tag k="building" v="yes"/>
</way>

The Portland extract has only one Node referenced from this Way, and that Node 4063882877 is not in the extract file.

OSM Way 404118908 is actually a closed Way object with 5 nodes denoting a car repair shop in Brazil. Here's what is looks like from the OSM API (still at version 1):

<way id="404118908" visible="true" version="1" changeset="37879771" timestamp="2016-03-16T18:44:38Z" user="Marcos Medeiros" uid="2607707">
<nd ref="4063882880"/>
<nd ref="4063882878"/>
<nd ref="4063882874"/>
<nd ref="4063882871"/>
<nd ref="4063882877"/>
<nd ref="4063882880"/>
<tag k="building" v="yes"/>
<tag k="name" v="Auto Elétrica Ituiutaba"/>
<tag k="phone" v="+55 34 3268-1246"/>
<tag k="shop" v="car_repair"/>
</way>

Neither this Way, nor any of the referenced Nodes should make their way into the Portland, Oregon extract as they should clearly be outside of the bounding box.

[buma]

There is also some nodes and ways from Germany in the extract from DC Baltimore

 $ grep 'id="149694"' dc-baltimore_maryland.osm_01.osm 
    <way id="149694" version="4" timestamp="2013-01-24T21:57:29Z" changeset="14775106" uid="5042" user="huskytreiber">

Way 149694. This in itself is strange but most problematic is that there are a lot of ways for which nodes doesn't exist in extract and it breaks my loading scripts.

Edit statistics: This is from workable extract seems from May last year

 osmconvert dc-baltimore_maryland.osm.pbf.OLD --out-statistics
timestamp min: 2005-11-30T04:26:09Z
timestamp max: 2015-05-30T01:53:55Z
lon min: -77.5989999
lon max: -76.0580010
lat min: 38.5390000
lat max: 39.6310000
nodes: 11715588
ways: 1314878
relations: 6459
node id min: 234661
node id max: 3557260796
way id min: 4415891
way id max: 349940888
relation id min: 218
relation id max: 5214556
keyval pairs max: 275
keyval pairs max object: relation 148838
noderefs max: 1970
noderefs max object: way 231269487
relrefs max: 741
relrefs max object: relation 4799101

And new one which doesn't work:

osmconvert dc-baltimore_maryland.osm.pbf --out-statistics
timestamp min: 2005-11-30T04:26:09Z
timestamp max: 2016-04-02T00:20:17Z
lon min: -77.5989999
lon max: -76.0580005
lat min: 38.5390000
lat max: 39.6310000
nodes: 12563456
ways: 1698343
relations: 22801
node id min: 234661
node id max: 4091227792
way id min: 531
way id max: 407085579
relation id min: 218
relation id max: 6096987
keyval pairs max: 321
keyval pairs max object: relation 52411
noderefs max: 1983
noderefs max object: way 356133005
relrefs max: 767
relrefs max object: relation 4799101

The most strange thing is that way ID min is smaller on new extract which is kinda impossible. And way with min ID 531 is in England UK.

[IndyHurt]

@dobratzp @buma this should be fixed via a prevent element leaks pull request. Check out the latest metro extracts available today (4/12/2016) and please let us know if the problem persists.

[buma]

This is definitely fixed for me in DC Baltimore extract. Thanks for the fix.