topic Zonalsum problem in ArcGIS Spatial Analyst Questions https://community.esri.com/t5/arcgis-spatial-analyst-questions/zonalsum-problem/m-p/738998#M10717 <HTML><HEAD></HEAD><BODY><SPAN>I have a binary grid that I want a zonalsum for. When I run zonal statistics in ArcMap Spatial Analyst using the menu option, the zonal statistics tool in ArcToolbox or in zonalsum in raster calculator I am getting the same incorrect results. I am getting large values in areas that are entirely 0 (should sum to 0). I have tried this with both the data and nodata flags. My zonal grid is 59.5MB (3.07GB uncompressed) (row/col dimensions are 32176/51272 at 90m with 41603 unique zonal values). The binary grids are the same dimensions but only ~26MB (1.7GB uncompressed). If I change the extent and subset down to the area that I am getting erroneous values for, I am getting the correct answer so it seems to be a large grid issue (again). It seems like ArcMap often yields incorrect answers with very large grids. I WOULD RATHER GET NO ANSWER THEN THE INCORRECT ONE! Unfortunately, this will not run in Workstation ArcInfo (I get a MSEEK error). </SPAN><BR /><BR /><SPAN>What I actually want is if the binary grid has &gt; 5 cells with a value of 1 in a zone assign 1 else 0.&nbsp; </SPAN><BR /><SPAN>CON(ZONALSUM(zonegrd, grd) &gt;= 5, 1, 0)</SPAN><BR /><BR /><SPAN>Since I am working with 1200+ binary grids, I would like to avoid breaking my data into smaller grids. Anybody have any suggested workarounds?</SPAN></BODY></HTML> Fri, 18 Jun 2010 00:35:42 GMT JeffreyEvans 2010-06-18T00:35:42Z Zonalsum problem https://community.esri.com/t5/arcgis-spatial-analyst-questions/zonalsum-problem/m-p/738998#M10717 <HTML><HEAD></HEAD><BODY><SPAN>I have a binary grid that I want a zonalsum for. When I run zonal statistics in ArcMap Spatial Analyst using the menu option, the zonal statistics tool in ArcToolbox or in zonalsum in raster calculator I am getting the same incorrect results. I am getting large values in areas that are entirely 0 (should sum to 0). I have tried this with both the data and nodata flags. My zonal grid is 59.5MB (3.07GB uncompressed) (row/col dimensions are 32176/51272 at 90m with 41603 unique zonal values). The binary grids are the same dimensions but only ~26MB (1.7GB uncompressed). If I change the extent and subset down to the area that I am getting erroneous values for, I am getting the correct answer so it seems to be a large grid issue (again). It seems like ArcMap often yields incorrect answers with very large grids. I WOULD RATHER GET NO ANSWER THEN THE INCORRECT ONE! Unfortunately, this will not run in Workstation ArcInfo (I get a MSEEK error). </SPAN><BR /><BR /><SPAN>What I actually want is if the binary grid has &gt; 5 cells with a value of 1 in a zone assign 1 else 0.&nbsp; </SPAN><BR /><SPAN>CON(ZONALSUM(zonegrd, grd) &gt;= 5, 1, 0)</SPAN><BR /><BR /><SPAN>Since I am working with 1200+ binary grids, I would like to avoid breaking my data into smaller grids. Anybody have any suggested workarounds?</SPAN></BODY></HTML> Fri, 18 Jun 2010 00:35:42 GMT https://community.esri.com/t5/arcgis-spatial-analyst-questions/zonalsum-problem/m-p/738998#M10717 JeffreyEvans 2010-06-18T00:35:42Z