Is this Comment-field the definitely format for this string - then in your example is no unique delimiter for the rows? If you have one it would be quite easy to split the string per subfield(string, delimiter) which split the string on the delimiter and repeated the other parts from this record. Quite similar is to use a while-loop whereby iterno() <= substringcount(delimiter). The other string-parts for the old- and new value could be split per subfield(string, delimiter, number) or textbetween() or maybe some nested mid/left/right- and index-functions.
I think if you haven't such unique delimiter you will need to do it quite similar only that you will need additionally steps within a if-loop where the current record-value with a previous record will be per peek() or previous() compared and the record-values will be adjusted. The worser the datastructure are the more efforts do you need with this comparing, checking and adjusting.
Maybe the following links are helpful in understanding what I mean: