Closed Seba95 closed 7 years ago
Hello Sébastien,
imho: you have to check: unique osm_id vs. unique osm_id
Because osm2pqsl
is splitting long ways : see https://help.openstreetmap.org/questions/40888/osm2pgsql-import-splits-ways-in-table-planet_osm_line
imho : only the select count(*)
is meaningless info ; Can you give a missing - real osm_id examples, with a links ? ( like this relation/3726175 )
and you have to check a relation and a member ids
like :
relation id="3726175"
) member type="way"
) <osm version="0.6" generator="CGImap 0.5.8 (15710 thorn-02.openstreetmap.org)" copyright="OpenStreetMap and contributors" attribution="http://www.openstreetmap.org/copyright" license="http://opendatacommons.org/licenses/odbl/1-0/">
<relation id="3726175" visible="true" version="5" changeset="41692003" timestamp="2016-08-25T15:57:02Z" user="Malvolapukulo" uid="4128818">
<member type="way" ref="228633464" role="outer"/>
<member type="way" ref="228633444" role="outer"/>
<member type="way" ref="228633442" role="outer"/>
<member type="way" ref="228633447" role="outer"/>
<member type="way" ref="228633451" role="outer"/>
<member type="node" ref="768032316" role="admin_centre"/>
<tag k="admin_level" v="4"/>
<tag k="boundary" v="administrative"/>
<tag k="ISO3166-2" v="EG-MN"/>
<tag k="name" v="محافظة الفيوم"/>
<tag k="name:de" v="Minya"/>
<tag k="name:en" v="Minya Governate"/>
<tag k="name:eo" v="Provinco Fajum"/>
<tag k="name:hu" v="el-Minja kormányzóság"/>
<tag k="name:pl" v="Muhafaza Fajum"/>
<tag k="name:ru" v="мухафаза Эль-Минья"/>
<tag k="type" v="boundary"/>
</relation>
</osm>
( disclaimer: I am only a power user )
Hi ImreSamu,
i really think that items are missing because the negative osm ids used to build the admin-*-zoom layers are missing.
I only see positive osm ids in the table built with imposm3.
In the osmcarto style, admin layers are built with this request:
(SELECT way, admin_level FROM planet_osm_roads WHERE boundary = 'administrative'
AND admin_level IN ('0', '1', '2', '3', '4')
AND osm_id < 0 <====== NEGATIVE ID
ORDER BY admin_level DESC) AS admin_low_zoom"
Regards, Sébastien
imho: this is not so simple ...
my advice, choose a very simple test case ...
"Loaded – nodes: 1941, ways: 44, relations: 1"
"Displayed – pois: 2, lines: 1, polygons: 1"
name
keyfor me sometimes helps in debugging - if I pre-filter the data ( avoiding noise )
imho:
I made a little test .. with /egypt-latest.osm.pbf ( I have added a wikipedia and a 'name:en'key to the admin mapping )
map=# select osm_id, name, name_en, wikipedia, admin_level, type from osm_admin order by admin_level asc ;
osm_id | name | name_en | wikipedia | admin_level | type
-----------+---------------------------------+----------------------------+--------------------------+-------------+----------------
317308949 | | | | 2 | administrative
-3061757 | محافظة أسوان | Aswan Governorate | | 4 | administrative
-3061846 | الإسكندرية | Alexandria Governorate | | 4 | administrative
-3062185 | الغربية | Suez Governorate | | 4 | administrative
-3062184 | الإسماعيلية | Ismailia Governorate | | 4 | administrative
-3061826 | مطروح | Matrouh Governorate | | 4 | administrative
-3061827 | الوادي الجديد | New Valley Governorate | | 4 | administrative
-3726124 | محافظة الفيوم | Faiyum Governorate | | 4 | administrative
-3726170 | محافظة بنى سويف | Beni Suef Governorate | | 4 | administrative
-3824206 | محافظة الجيزة | Giza Governorate | | 4 | administrative
-3824513 | البحيرة | Beheira Governorate | | 4 | administrative
-3824207 | محافظة المنوفية | Monufia Governorate | | 4 | administrative
-4103336 | القاهرة | Cairo Governorate | | 4 | administrative
-4103404 | دمياط | Damietta Governorate | | 4 | administrative
-4103405 | كفر الشيخ | Kafr el-Sheikh Governorate | | 4 | administrative
-4103403 | الدقهلية | Dakahlia Governorate | | 4 | administrative
-4103337 | القليوبية | Qalyubia Governorate | | 4 | administrative
-4103406 | محافظة بورسعيد | Port Said Governorate | | 4 | administrative
-4103407 | الشرقية | Al Sharqia Governorate | | 4 | administrative
-3584607 | محافظة الغربي | Gharbiyya Governorate | | 4 | administrative
-3060793 | جنوب سيناء | South Sinai Governorate | | 4 | administrative
-3726175 | محافظة الفيوم | Minya Governate | | 4 | administrative
-3726184 | أسيوط | Asyut Governorate | | 4 | administrative
-3726186 | سوهاج | Sohag Governorate | | 4 | administrative
-3726189 | محافظة قنا | Qena Governorate | | 4 | administrative
-3726211 | الأقصر / Luxor | Luxor Governate | de:Gouvernement al-Uqsur | 4 | administrative
-3586570 | محافظة الغربية | | | 7 | administrative
-3585231 | مديرية الغربية | | | 7 | administrative
-3603506 | طنطا | Tanta | | 8 | administrative
-3608214 | برما | Birma | | 8 | administrative
-3608407 | كفر شعبان | Kafr Shaban | | 9 | administrative
-3587035 | ميت السودان | | | 9 | administrative
-3605484 | شوبر | Shobar | | 9 | administrative
-3605393 | كفر عصام | Kafr Isam | | 9 | administrative
-3588686 | بريك الحجر | Bureik al Hagar | | 9 | administrative
-3578930 | سبرباي | | | 9 | administrative
-4081799 | وادي حلفا | Wadi Halfa Salient | | | administrative
(37 rows)
As I see the country border is missing :(
imho: As I see the Geofabric extract poly for Egypt is smaller than the real border (admine_level=2 ) https://gist.github.com/anonymous/b2de797a46446cd2be29bd388ec47935
So the extract only contains - part of the country border / part of Egypt osm data ;
so this is a reason why not included the admin_level=2 in the impom3 type: polygon
mappings .. :disappointed:
imho : We have one more problems.. There is a data problem here ..
As I remember - If you want to compare the osm2pgsql with the imposm3
it is better to add a--multi-geometry (-G)
flag to the osm2pgsql ;
see : https://github.com/openstreetmap/osm2pgsql/blob/master/docs/analysis.md
Hi ImreSamu,
thx for the info about multi-geometry it has indeed the same pb. Are there other kind of objects with that issue? It would be nice to have a non multi-geometry option in imposm3 to be able to render things like osm2pgsql even if the data is not perfect.
Regards, Sébastien
imho:
If you want compare, than you have to modify the osm2pgsql line splitting : https://github.com/openstreetmap/osm2pgsql/search?utf8=%E2%9C%93&q=Split+long+ways+after+around+1+degree+or+100km And on the imposm3 side: you can use PostGIS post-processing - to split or merge lines ...
But the PostGIS post-processing always works ..
Hi,
i am using this file to test imposm3 : http://download.geofabrik.de/africa/egypt-latest.osm.pbf with the example-mapping.json.
Could you confirm the issue?
Regards, Sébastien