[OSM-dev] Deleting TIGER node tags

[Matt Amos]

On Sat, Jun 27, 2009 at 2:28 PM, Thomas Wood<grand.edgemaster at gmail.com> wrote:
> 2009/6/27 Ed Loach <ed at loach.me.uk>:
>>> Indeed.  Leave it in the ways, remove it from the nodes.
>>
>> Is this a low (database) level delete? Or will we gain a new history
>> entry for every node showing the node as it was before the delete
>> and as it was after? If the latter then won't the database end up
>> bigger rather than smaller?
>
> The database is a matter for the sysadmins, this will however impact
> greatly on the planetfile size, which more people often struggle with.

i think a database-level delete isn't necessary. it does mean the
current_node_tags table will shrink - maybe by 30% or so, but the
nodes table will grow. overall the DB might stay roughly the same size
(more node versions, but less tag duplication in the current_node_tag
tables).

doing a "proper" delete via the API also means the correct osmchange
diffs will be generated, so anyone using those for replication won't
have to worry about it.

as thomas says, the planet will shrink. also, map calls will be
smaller in the US. this means faster data processing and more
responsive editing for everyone. :-)

cheers,

matt

PS: although it would be fantastic if, in any future imports, the
number of tags on every element were kept to the absolute minimum.
e.g: put source tags in the changeset, not on the element; don't
duplicate way tags onto the node; don't put tags on elements to
describe geometry (tiger:county tags are utterly pointless, as the
county boundaries are in OSM already, meaning the county an element is
in can be determined by a simple geometric calculation). </rant>