id summary reporter owner description type status priority milestone component version resolution keywords cc 3318 ST_MapAlgebra with mask seems to hard-code distance 1 instead of inferring distance from mask matrix robe Bborie Park "I've moved this from the now closed mask documentation #2603 to here, I'm still having issue with this even as of r14198 even though dusty has made some changes. {{{ DROP TABLE IF EXISTS shapes; CREATE TABLE shapes(rid integer, rast raster); INSERT INTO shapes(rid,rast) VALUES ( 1, ST_AsRaster( ST_Buffer( ST_GeomFromText('LINESTRING(50 50,150 150,150 50)'), 10,'join=bevel'), 200,200,ARRAY['8BUI'], ARRAY[118], ARRAY[0]) ); SELECT st_mapalgebra(rast,1,'st_mean4ma(double precision[], int[], text[])'::regprocedure,'{{1,0,1,1,1}, {1,0,1,1,1}, {1,0,1,1,1}, {1,0,1,1,1}, {1,0,1,1,1}}'::double precision[],false) FROM shapes; }}} Gives error: {{{ ERROR: rt_pixel_set_array: mask dimensions 5 x 5 do not match given dims 3 x 3 }}} It's my understanding based on dusty's comments that the neighborhood size should be inferred from the mask size, so there seems to me something wrong here. Bborie's comments copied from other ticket ---- Distance is two pixels left, right, up, down from the pixel of interest... {{{ 2 2 2 2 2 2 1 1 1 2 2 1 0 1 2 2 1 1 1 2 2 2 2 2 2 }}} So the distance x/y are correct. As for that error message, I'll need to dig. ---- " defect assigned medium PostGIS Fund Me raster master