Investigate all dynamically set sections of the cdf file when writing

[bryan-harter]

In PR #287 I had to make a change to the creation of an aedr. In it, is used to determine the size of the AEDR by calculating the length of the string. This led to issues where the length of the data object we were writing did not match the expectations one would have when reading the file. So I changed it to:

value_size = recs * self._datatype_size(dataType, numElems)

from

value_size = len(cdata)

Are there other locations in the code where this problem also arises?