rf_rasterize of circles of same diameter return tiles of different sizes

[jornfranke]

This might not be an issue, but I could not verify this understanding in the documentation. Here is what I do:

Input: Raster "A" (cells with continuous values) Multiple Points (Geometry)

Process:

• Create a 1000m buffer (circle) around each point using st_bufferPoint
• Join (intersect) Raster "A" and the circles
• rasterize each circle to create a mask
• apply the mask to the Raster "A"
• Calculate statistics (rf_agg_stats) on the masked raster.

The statistics (min,max,mean,variance) look good and plausible. I have only one issue with data_cells and na_datacells. For some masks the total sum of the two is 65536 and others it is 131072 (=2*65536). Is this correct? I would have expected it would always be the same. Maybe if the circle falls at certain positions on the raster mask there is a need to make it double the size. As said, the other statistics are correctly calculated.