xgunnerx ... this isn't an option because of the following real life situation:

Two atomic entries found in my data set:

N'-phosphoguanidinoethyl methyl phosphate
2N N'-phosphoguanidinoethyl methyl phosphate
[download]

For the first entry your script will return the following when it should return an empty string as there is no prefix:

N'-phosphoguanidinoethyl
methyl
[download]

For the second entry it will return the following when it should return "2N" as it is the prefix (note that the entire name minus prefix is not returned by your code as the 'methyl phosphate' biut is missing:

2N
N'-phosphoguanidinoethyl
[download]

I admit that given the original data set that I provided your solution would work, but it does not allow for arbitrary names with spaces.

Arun