Re: [OSM-talk] Relation #2632934 is "killing" differential update

[mmd]

On 2/11/19 11:41 AM, wambac...@posteo.de wrote:
> Hi,
> 
> there is a relation https://openstreetmap.org/relation/2632934, which
> currently has *42698 members*. Processing this rel with osm2pgsql fails,
> because this value is bigger than 32737 (small int).
> 

FYI: As a follow up to this discussion, the new cgimap based changeset
upload will introduce a limit of 32'000 relation members. This seems
like a good choice for many tools in the OSM ecosystem, including osm2pgsql.

See https://github.com/zerebubuth/openstreetmap-cgimap/pull/174 for
further details and discussion.

-- 



___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Relation #2632934 is "killing" differential update

[Roland Olbricht]

There must have happened something ugly at versions 99 and 106:
https://overpass-turbo.eu/s/FYm


Thanks for that - yet again something useful in overpass-turbo ("make 
stat") that I had no idea existed.  Is there any documentation about it 
anywhere?  There is 
https://wiki.openstreetmap.org/wiki/Overpass_API/Overpass_QL#The_statement_make 
, but that doesn't have any useful information or give any examples.


Thank you for the feedback. The idea is laid out in
http://dev.overpass-api.de/blog/sliced_time_and_space.html
in particular
http://dev.overpass-api.de/blog/sliced_time_and_space.html#timeline
within
http://dev.overpass-api.de/blog/sliced_time_and_space.html#lrs

However, I accept that it is not so obvious how to combine the two. I'm 
sorry for all that have missed that. I have added an example on

https://wiki.openstreetmap.org/wiki/Overpass_API/Overpass_API_by_Example#Making_a_Table_of_Object_Versions

Best regards,

Roland

___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Relation #2632934 is "killing" differential update

[wambacher]

it's an old problem: https://github.com/openstreetmap/osm2pgsql/pull/716

need to check my osm2pgsql  (is 0.93 and should be ok, but it is not)

walter


-- 
My projects:

Admin Boundaries of the World 
Missing Boundaries

Emergency Map 
Postal Code Map (Germany only) 
Fools (QA for zipcodes in Germany) 
Postcode Boundaries of Germany 
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Relation #2632934 is "killing" differential update

[ajt1...@gmail.com]

On 11/02/2019 10:59, Roland Olbricht wrote:


There must have happened something ugly at versions 99 and 106:
https://overpass-turbo.eu/s/FYm


Thanks for that - yet again something useful in overpass-turbo ("make 
stat") that I had no idea existed.  Is there any documentation about it 
anywhere?  There is 
https://wiki.openstreetmap.org/wiki/Overpass_API/Overpass_QL#The_statement_make 
, but that doesn't have any useful information or give any examples.


Best Regards,

Andy






___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Relation #2632934 is "killing" differential update

[Jo]

I split the route relation in 2. Didn't fix the erroneous use of
route_master though.

Polyglot

On Mon, Feb 11, 2019 at 12:30 PM Jo  wrote:

> I'm looking into it, but someone completely misunderstood the use of
> route_master relations and all those relations should be split in at least
> 2, for each direction of travel / variation in itinerary.
>
> Jo
>
> On Mon, Feb 11, 2019 at 12:27 PM Maarten Deen  wrote:
>
>> On 2019-02-11 12:14, Roland Olbricht wrote:
>> > Hi,
>> >
>> > the relevant changesets are
>> > 65448087 (for version 99, created_by "upload.py v. 1") and
>> > 67093992 (for version 106, created_by "Vespucci 12.1.2.0").
>> >
>> > I cite the user agents because the data looks like unintentionally
>> > changed that way.
>> > I have left a bug report on Vespucci, see
>> > https://github.com/MarcusWolschon/osmeditor4android/issues/859
>>
>> I have removed all duplicates from the relation but done nothing to
>> check remaining data for correctness. It's only 350 unique members.
>>
>> Maarten
>>
>> ___
>> talk mailing list
>> talk@openstreetmap.org
>> https://lists.openstreetmap.org/listinfo/talk
>>
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Relation #2632934 is "killing" differential update

[Jo]

I'm looking into it, but someone completely misunderstood the use of
route_master relations and all those relations should be split in at least
2, for each direction of travel / variation in itinerary.

Jo

On Mon, Feb 11, 2019 at 12:27 PM Maarten Deen  wrote:

> On 2019-02-11 12:14, Roland Olbricht wrote:
> > Hi,
> >
> > the relevant changesets are
> > 65448087 (for version 99, created_by "upload.py v. 1") and
> > 67093992 (for version 106, created_by "Vespucci 12.1.2.0").
> >
> > I cite the user agents because the data looks like unintentionally
> > changed that way.
> > I have left a bug report on Vespucci, see
> > https://github.com/MarcusWolschon/osmeditor4android/issues/859
>
> I have removed all duplicates from the relation but done nothing to
> check remaining data for correctness. It's only 350 unique members.
>
> Maarten
>
> ___
> talk mailing list
> talk@openstreetmap.org
> https://lists.openstreetmap.org/listinfo/talk
>
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Relation #2632934 is "killing" differential update

[Maarten Deen]

On 2019-02-11 12:14, Roland Olbricht wrote:

Hi,

the relevant changesets are
65448087 (for version 99, created_by "upload.py v. 1") and
67093992 (for version 106, created_by "Vespucci 12.1.2.0").

I cite the user agents because the data looks like unintentionally
changed that way.
I have left a bug report on Vespucci, see
https://github.com/MarcusWolschon/osmeditor4android/issues/859


I have removed all duplicates from the relation but done nothing to 
check remaining data for correctness. It's only 350 unique members.


Maarten

___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Relation #2632934 is "killing" differential update

[Roland Olbricht]

Hi,

the relevant changesets are
65448087 (for version 99, created_by "upload.py v. 1") and
67093992 (for version 106, created_by "Vespucci 12.1.2.0").

I cite the user agents because the data looks like unintentionally changed that 
way.
I have left a bug report on Vespucci, see
https://github.com/MarcusWolschon/osmeditor4android/issues/859

I have not found a handle for "upload.py".

Best regards,

Roland

___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Relation #2632934 is "killing" differential update

[wambacher]

Hi

> this is an Osmosis problem.

oops, osmosis.

running osmosis 0.45. has this problem (int2 -> int4) solved in any
other version?

> Nonetheless, the relation does not make sense. It is an ordinary bus route in 
> Manhattan, but members hundreds of times repeated.
>
> There must have happened something ugly at versions 99 and 106:
> https://overpass-turbo.eu/s/FYm

Yes, 106 is the reason.

what will happen now? if it's a problem, which has been fixed in newer
versions of osmosis, it's my problem.

but what if not? "somebody" has to remove this update from the osc-files.

regards

walter

-- 
My projects:

Admin Boundaries of the World 
Missing Boundaries

Emergency Map 
Postal Code Map (Germany only) 
Fools (QA for zipcodes in Germany) 
Postcode Boundaries of Germany 
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

Re: [OSM-talk] Relation #2632934 is "killing" differential update

[Roland Olbricht]

Hi,

> there is a relation https://openstreetmap.org/relation/2632934, which 
> currently has 42698 members. Processing this rel with osm2pgsql fails, 
> because this value is bigger than 32737 (small int).

this is an Osmosis problem.

Nonetheless, the relation does not make sense. It is an ordinary bus route in 
Manhattan, but members hundreds of times repeated.

There must have happened something ugly at versions 99 and 106:
https://overpass-turbo.eu/s/FYm

Best regards,
Roland

___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk

[OSM-talk] Relation #2632934 is "killing" differential update

[wambacher]

Hi,

there is a relation https://openstreetmap.org/relation/2632934, which
currently has *42698 members*. Processing this rel with osm2pgsql fails,
because this value is bigger than 32737 (small int).

Reading in file: /osm/db/planet3/diffs/2019/02/11/2019-02-11T10:33:01.osc
Using XML parser.
node cache: stored: 1591(100.00%), storage efficiency: 17.37% (dense
blocks: 1, sparse nodes: 484), hit rate: 28.31%
Osm2pgsql failed due to ERROR: insert_rel failed: FEHLER:  Wert »42698«
ist außerhalb des gültigen Bereichs für Typ smallint
(7)
Arguments were: 2632934, 4616, 42698, {6116226169,..

Is this a general or an osmosis problem? I'm running still osmosis 0.45.

OSC: https://planet.openstreetmap.org/replication/minute/003/362/071.osc.gz

State:

#Mon Feb 11 09:32:04 UTC 2019
sequenceNumber=3362071
txnMaxQueried=1978118272
txnActiveList=1978118258
txnReadyList=
txnMax=1978118272
timestamp=2019-02-11T09\:32\:02Z

Regards
walter



-- 
My projects:

Admin Boundaries of the World 
Missing Boundaries

Emergency Map 
Postal Code Map (Germany only) 
Fools (QA for zipcodes in Germany) 
Postcode Boundaries of Germany 
___
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk