The difference of transforming raster file in different cordinates between export and projection

Question

I am using ArcGIS Desktop version 10.4.1. I have a raster file with Geographic coordinate system WGS84. For my analysis, I needed to transform it into UTM projection. So I used the export function: http://desktop.arcgis.com/en/arcmap/10.3/manage-data/raster-and-images/exporting-a-raster-in-arcmap.htm

I used The Data Frame (Current) option t export the raster dataset using the spatial referencing specifications of the data frame. So the coordinate of the current data frame (UTM) is different from the coordinate of the raster file. It works very well.

However, when I use the project function of ArcMap to re-project from geographic coordinate to UTM and conducted the analysis, I noticed that the results are somehow different.

I googled all possible documents explaining the difference of export and project, I could not find a document explain the difference of those function

Can anybody explain the difference of two functions?

It seems that export is doing the same thing as project when the coordinates of the source and the current data frame are different. But I want to make sure.

Answer 1

After more extensive search, especially, ArcGIS bug reports, I found that ArcGIS has a bug when one uses the raster projection from geographic coordinate to projection coordinate and when the cell is blank. https://support.esri.com/en/technical-article/000007126

To solve this problem, they suggest to use export instead of raster projection. I think that this is the reason I got the different results using export and projection.