> It needs to be "way-based" because only then the updates will work.
OK, I had assumed that both way and node geometries needed to be in
`highways_end` so that both updates that removed the node or way would work.
> From a performance standpoint I would recommend to use several small tables
> instead of a few large tables
It's not really significant, since any change to what is included as "nodes in
ways" will require a re-import. It would probably be more elegant to use a
table per feature anyway.
Just to put some numbers in:
`highway=turning_circle`: 2.5m
`highway=level_crossing`: 997k
`ford=X`: 1.4m (nodes)
[`railway=station` won't work because these are not generally nodes on ways.]
Does that seems manageable?
But it does sound like the next stage would be for me to see how things work in
practice on real data. May be some time, as I haven't tried compiling
`osm2pgsql` from source...
--
Reply to this email directly or view it on GitHub:
https://github.com/osm2pgsql-dev/osm2pgsql/discussions/2455#discussioncomment-16546916
You are receiving this because you are subscribed to this thread.
Message ID:
<osm2pgsql-dev/osm2pgsql/repo-discussions/2455/comments/[email protected]>
_______________________________________________
Tile-serving mailing list
[email protected]
https://lists.openstreetmap.org/listinfo/tile-serving
