Skip to main content
I have an Esri Raster that is supposed to be an elevation model, but I can't persuade FME to recognize it has Z values with NoData. The water features are nodata and it seems that I cannot get rid of the extremely low values so that it draws as a nice surface. The only DEM example I can find uses a CDED format which is a bit obscure - even if it is from the Home of FME!

 

I eventually found RasterCellCoercer and tested for Z values > 0. A RasterReSampler first kept the data volumes acceptable. I was looking for a lattice type poeration, but I suppose points with explicit X,Y coordinates will do.

 


Hi Kim,

 

 

I think the RasterCellCoercer will ignore automatically NoData cells when you set 'No' to 'Extract Nodata Values' parameter.

 

 

Takashi
One would expect that, but I had to add a test. FME does not seem to recognise the value as NoData (1.0E-38). I also tried to use RasterBandNoDataRemover, clip the raster to remove the sea areas, and a whole lot of other blind alleys.

 


Maybe it is the way I have clipped the data. The original has valid NoData, but  not after clipping.The NoData is suddenly treated as normal values.

 

I will backtrack to see what happened there.

 


Hi again,

 

 

Are you using RasterBandNodataRemover in your workflow? RasterBandNodataRemover removes Nodata value definition rather than Nodata cells. So after applying this transformer, any other transformers like RasterCellCoercer will not be able to recognize ex-Nodata cells.

 

 

Takashi
Hi,

 

 

Maybe this answer of Mark can shed some light.

 

Hope this helps.

Reply