Steve Ratcliffe wrote:
Was there anything special about what did end up in the file? I tried it out and although my file was not big enough to need splitting, all of it ended up in the output file.
The nodes and ways in the resulting 63240001.osm.gz look fine to me. The only difference I see is that srtm2osm uses double quotation marks and splitter.jar uses single quotation marks.
Were there any messages? It could be that contour lines are very long and regularly cross more than four tiles.
I made another test with a smaller file and this is what it says: $java -Xmx2048M -jar splitter.jar --mapid=31000001 --max-nodes=100000 srtm_5.8_47.7.osm key mapid/ val 31000001 key max-nodes/ val 100000 comparing size 6 a (47.9443359375,6.0205078125) to (47.98828125,6.064453125), size=6 Total time 10s The result is: -rw-r--r-- 1 unrz96 unrz 99M 2009-01-29 09:52 srtm_5.8_47.7.osm -rw-r--r-- 1 unrz96 unrz 158 2009-01-29 09:52 areas.list -rw-r--r-- 1 unrz96 unrz 340 2009-01-29 09:53 template.args -rw-r--r-- 1 unrz96 unrz 535K 2009-01-29 09:53 63240001.osm.gz The original file had ~ 1 million nodes: $grep "<node id" srtm_5.8_47.7.osm | wc -l 1041479 The result has just 15086: $gzip -cd 63240001.osm.gz | grep "<node id" | wc -l 15086 BTW: The name of the output file is not correct, I had set the map id to 31000001 but it used 63240001.
Otherwise I have no idea, could you put up the files you were using somewhere?
Sure, here it is: http://www.kleineisel.de/ralf/gps/splitter/