Hi coverage friends,
I have a strange scaling issue that is giving troubles to my testcases
and need some advice.
Let's assume I have a small raster containing the data:
{1000.0, 1000.0, 1200.0, 1250.0, 1300.0, 1350.0, 1450.0}, //
{750.0, 850.0, 860.0, 900.0, 1000.0, 1200.0, 1250.0}, //
{700.0, 750.0, 800.0, 850.0, 900.0, 1000.0, 1100.0}, //
{650.0, 700.0, 750.0, 800.0, 850.0, 490.0, 450.0}, //
{430.0, 500.0, 600.0, 700.0, 800.0, 500.0, 450.0}, //
{700.0, 750.0, 760.0, 770.0, 850.0, 1000.0, 1150.0} //
I use the jai scxaling operation to scale it from 6 rows to be 5 rows.
Well, on linux and mac systems I get:
{1000.0, 1200.0, 1250.0, 1300.0, 1450.0}, //
{750.0, 860.0, 900.0, 1000.0, 1250.0}, //
{700.0, 800.0, 850.0, 900.0, 1100.0}, //
{430.0, 600.0, 700.0, 800.0, 450.0}, //
{700.0, 760.0, 770.0, 850.0, 1150.0} //
While on windows I get:
{1000.0, 1200.0, 1250.0, 1300.0, 1450.0}, //
{750.0, 860.0, 900.0, 1000.0, 1250.0}, //
{650.0, 750.0, 800.0, 850.0, 450.0}, //
{430.0, 600.0, 700.0, 800.0, 450.0}, //
{700.0, 760.0, 770.0, 850.0, 1150.0} //
The fact is that the center of the resulting 3. row is between row 3
and 4 of the starting data. Probably some floating point precision
thing?
In the case of nearest neighbor scaling, I assume the upper row should
be picked?
Any idea how to solve this. My latest banging the head against the
wall didn't give the necessary results.
Thanks,
Andrea
------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network
management toolset available today. Delivers lowest initial
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd
_______________________________________________
Geotools-devel mailing list
Geotools-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geotools-devel
