Thanks for the link Will. I don't quite understand what this
command is doing (UNIX pipes still confuse me). Is it taking the
two mbtiles files and outputting them to a new file .dump? I guess
there would need to be no overlap for this to work?
This command will merge tilesA.mdtiles into tilesB.mbtiles by
basically dumping the entire tilesA database to a SQL text file,
and feed that into tilesB.
You don't need to worry about errors like table X already
exists or index X already exists since it is
just trying to create tables that already exist in the destination
mbtiles. But you do need to worry about errors like column X
is no unique since this signifies there is a conflict
somewhere. You can check what is throwing the error by outputting
the SQL from tilesA.mbtiles to a text file and inspecting it:
Maybe someone from Mapbox can clarify this: looking through the
code for node-mbtiles it looks like tile_id is created as a hash
from image data, therefore any conflict of ids when merging two
mbtiles is because the image is the same (I imagine plain all one
color tiles) so that this conflict would not be something to worry
about. Is that correct?