[OSM-talk] Tag statistics

Lester Caine lester at lsces.co.uk
Wed Mar 7 08:11:42 GMT 2007


Robert (Jamie) Munro wrote:

> Lester Caine wrote:
>> I've just finally scanned last weeks planet file :)
>> Main problem was that I ha to switch off the foreign key constraints 
>> since there are segments without nodes and then ways without segments.
>>
>> So we do need something to go through and tidy the raw data.
>>
>> I'm running all this on a couple of slowish machines so its taking a few 
>> hours each time before problems crop up, but I'm about to re-run with 
>> this weeks file on a faster setup that I've been putting together in the 
>> meantime.
>>
>> Once I have the whole lot in Firebird I can start looking at using the 
>> data for searching :)
> 
> I'm a bit out of date with Firebird. Does it do R-Tree indexes yet, or
> is it going to?

Out on site yesterday ( 400 miles of tracks to upload ;) )

Firebird version 2 has made some major changes to the index structure, 
and while r-tree is not provided, the new structure should handle things 
a lot faster than the older simple b-tree structure. One of the reasons 
for wanting to get the data loaded is to check out the performance 
against both FB1.5 and other options.

Only problem I can see at the moment is how to keep thing in sync, since 
uploading the whole planet.osm every time seems a little drastic. While 
I can simply include records that have a created date later than that of 
the previous file, how do I monitor changes to older records? Is 
'created' the original created date or is it updated when a change is 
posted. A 'last update' date would be useful for sites which are slaving 
data, you would check the 'last updated' date for a way centrally, and 
then either download or use the local cache.

-- 
Lester Caine - G8HFL
-----------------------------
L.S.Caine Electronic Services - http://home.lsces.co.uk
Model Engineers Digital Workshop - 
http://home.lsces.co.uk/ModelEngineersDigitalWorkshop/
Treasurer - Firebird Foundation Inc. - http://www.firebirdsql.org/index.php




More information about the talk mailing list