Difference between revisions of "Shptree notes"

From OSGeo
Jump to navigation Jump to search
Line 1: Line 1:
 +
''(shptree processing notes for the [[Benchmarking_2010]] exercise)''
 +
 
== shptree Depth ==
 
== shptree Depth ==
  

Revision as of 10:52, 24 July 2010

(shptree processing notes for the Benchmarking_2010 exercise)

shptree Depth

No depth parameter was used for the shptree command, which means the shptree code would determine the optimal depth. The determined depths were:

file depth
building 21
contour-0 16
contour-1 17
contour-2 17
contour-3 17
contour-4 17
contour-5 15
contour-6 17
contour-7 17
industry 17
motorway 10
point-labels-for-geometry 15
point-labels-no-geometry 18
ramp 14
road 14
settlement 16
track 16

Verifying the qix file

  • shptreetst utility was used to verify the integrity of qix file
  • you should see a message like:
 $ shptreetst contour-0.qix 
 This new LSB index supports a shapefile with 221671 shapes, 16 depth 
 shapes 0, node 4, -8.187613,36.166759,3.274251,43.666687 
 shapes 0, node 4, -8.187613,36.166759,-1.883588,40.291719 
 shapes 0, node 4, -8.187613,36.166759,-4.720399,38.435487 
 shapes 0, node 1, -8.187613,36.166759,-6.280646,37.414560 
 shapes 0, node 2, -7.329478,36.728269,-6.280646,37.414560 
 ... 
 
 read entire file now at quad box rec 20 file pos 1000
 using last read box as a search 
 result of rectangle search was 
   86554, 86580, 86586, 86598, 86638, 86641, 86642, 86643, 86660, 86667, 195862, 195895, 195901, 195903, 195926, 196873, 197711,

buildings.shp

  • shptreetst building.qix caused a Segmentation fault initially on our Linux server (with 8GB RAM)
  • to solve this we...