Re^2: Speeding up point-in-polygon -- take two

[punkish]

If this is a task you're going to do repeatedly, and your polygon list is relatively static, then you should consider redefining the problem. I suspect that the loop and indexing operations my be consuming a considerable amount of your time. So maybe you can change the problem to unroll the loops.

Yes, the polys are relatively static, although the points are not so. Nevertheless, I do indeed unroll both polys and points, and move them into the SQLite db. That way I avoid going back to the files repeatedly.

You are correct about "loop and indexing operations my be consuming a considerable amount of your time". Below is the output from my latest run of 5000 polys using Devel::Profiler. More than 220k points were evaluated and updated, and my point-in-poly test was performed 220k times.

Total Elapsed Time = 251.1784 Seconds            
  User+System Time = 189.3034 Seconds            
Exclusive Times            
%Time  ExclSec  CumulS  #Calls  sec/call    Csec/c    Name
72.3   136.9    136.91  220726  0.0006      0.0006  main::_pointIsInPolygon
9.09    17.21    17.216   5001  0.0034      0.0034  DBI::st::fetchall_arrayref
6.25    11.82    11.825  13887  0.0009      0.0009  DBI::st::execute
4.84     9.162  178.99       1  9.1622    178.99    main::updatePoints
1.82     3.442    3.442   3885  0.0009      0.0009  DBI::db::commit
0.16     0.298    0.298   5001  0.0001      0.0001  DBI::st::fetchrow_array
0.07     0.141    0.141   3887  0           0       DBI::st::finish
0.01     0.01     0.01       3  0.0033      0.0033  IO::Handle::new
0        0        0.01       5  0           0.002   Geo::ShapeFile::get_bytes
0        0        0          6  0           0       DBI::_new_handle
0        0        0.01       5  0           0.002   Geo::ShapeFile::get_handle
0        0        0          1  0           0       DBI::connect
0        0        0          4  0           0       DBI::db::prepare
0        0        0          4  0           0       DBI::_new_sth
0        0        0          3  0           0       Geo::ShapeFile::dbf_handle  

What I really have to figure out is to reduce those tests. I have an idea which I have to now figure out how to implement. Basically, it is like so --

Once I get a bunch of potential points within the rect of a poly, I should find the 4 outermost points in each dimension. Once I determine those four definitively, all points within those 4 would be also within that poly, and I wouldn't have to do the checks for them.

Now the hard part -- realize the above hypothesis.