For a project I worked on last year, we were given spreadsheets of
legacy data by the
client. We used them to generate their own data load scripts by writing
spreadsheet
formulas to concatenate the values of certain columns into the correct
sql syntax in
another column. Then you just copy an entire column of generated inserts
into a text file.
We used open office since not everyone had a license for excel and we
found a) there are
some syntax differences between excel’s forumlas and those of open
office, and b) open
office gets very slow and unstable when you start apply complicated
concatenation formulas
across mulitple worksheets and thousands of rows! 
By the end, my co-worker who did the biggest data sets swore that he’d
use perl next
time… and well, we all know that anything perl can do ruby can do
better.
b
PS: also, open office can open most any excel file and save it as an
open office
spreadsheet file… which is… (drumroll) a zipped up collection of xml
files! So, you
could unzip the oo spreadsheet and get to work with ruby or xslt or
whatever. (I’ve heard
that MS’ latest version of office is supposed to use xml-based file
formats, but don’t
know if that’s out or what version you’re using…)