If the problem is what I suspect, I can explain the fact that some fields are all together.    If the entire location of the field would be within a decimal minute between .10 and .19, it would not be split.   But if  the same size field fell between say .18 and .23, then remembering that you loose the first digit after the decimal point, you would have the part of the field from .18 to .19 in one area of the map, and the part of the field between .20 and .23 in another area of the map (actually, lower or to the right), so these chunks would be reversed relative to each other.

As to the missing data... you are looking at two kinds of maps.   The old software is generating a "contour" map.  You are not seeing a map of individual sites.  You are seeing a map of estimated yield based upon the sites that got collected.    In the new software, you are seeing a dot at each logged site.   If you aske the new software to generate a surface, and theme it the same, it would probably look the same.