global data in extracts for Beirut and Lesbos

[geoBWM]

I created a new export this morning for data from Beirut (by drawing a polygon). I was very surprised to discover there is global data in the output. Something went wrong somewhere. PLease check. The export: https://osmaxx.hsr.ch/media/osmaxx/outputfiles/d958469e-a9d9-4541-8722-4f6847627a4d/lbn_beirut_wgs-84_2020-08-05_fgdb_full-detail.zip

Thanks

[das-g]

While it's normal and expected that OSMaxx extracts contain data outside the selected area, because features that are even just partially contained in that area are included in full, the linked extract contains objects from all over the world which indeed seems strange:

[geoBWM]

Once again, the problem of data from different parts of the globe being erroneously included in data extracts from OSMaxx has occured. See Extraction Order #8528 "Lesbos, Greece" From 34 layers produced, 31 of them contain data from locations clearly outwith the defined extent (only the top 3 are ok and have OK in their name).

There would seem to be a serious issue somewhere in the pipeline and it is making the service difficult to trust. Is this going to be investigated? Thanks

[das-g]

Is this going to be investigated?

Unfortunately probably not, at least not by geometalab staff. We are looking into replacing OSMaxx with (yet to be developed) simpler and easier-to-maintain software, that would be able to do the schema transformation OSMaxx did but which wouldn't perform the area extraction itself. (So one would need to feed it an excerpt obtained from another extraction service, such as https://export.hotosm.org/, or with static extraction results, as are available from https://download.geofabrik.de/.)

However, if anyone wants to look into the issue, we're open for any finding and would gladly take Pull Requests that tackle the problem.