Change Coodinate to byte[] or double[]

[13661172102]

For large geometry, there is a lot of points, then a lot of Coodinate Object. It takes multiple memory than byte[] or double[].

I know use Coodinate make the code readable. But i think performance is more important in future, because the spatial data size increase very fast.

[dr-jts]

This is why PackedCoordinateSequence exists.

[13661172102]

thanks u , it means a lot. I will try to extend PackedCoordinateSequence for my own @dr-jts

[13661172102]

@dr-jts i have tried to use PackedCoordinateSequence . But I found that when i do something like intersection ， the toCoordinateArray() of PackedCoordinateSequence will be called. Then double[] turns to Coordinate[] again and the memory use is largger than before because double[] and Coordinate[] both exists.

[dr-jts]

The Coordinate[] should only exist while the computation is happening, and then be freed.