Rick den Haan noted: >>For reference, I tried changing the character set, as Ron suggested, to a lot >>of different types, but all of them gave me various incarnations of gibberish. >>unpack() appears to solve my problem. Hi Rick, I'm glad you are making progress. ;-) Sorry, I thought your original note said that you had a 'VERY' detailed file spec. I assumed that you meant that the spec already told you which fields were 'zone packed decimal' or 'multi-punch column binary' - which, if this data is really *that* old are format types you could easily run into. If you are going to have to _guess_ at each of the field types then I think you are in for some very long days.  Anyway here is a link that may help as a starting point to enjoying old BCD and COBOL "comp-3" fields and how they can differ from old IBM signed fields: http://www.3480-3590-data-conversion.com/article-packed-fields.html It also throws a number of keywords at you that you can google for more info. Good Luck! RonL.  Especially if you don't know the underlying system the data came from. I would check around to be sure but I would not assume that a 'packed' field pulled off an old Univac machine could be properly decoded with an 'unpack' function pulled off an old Honeywell machine. I think there might be a potential for vendor specific implementations.