Have you done a rollback of the affected database table after the failed attempt(s)? If not, then of course the same old input data file will now have no suitable records to work on in the table, because those old "ccode" values are not present anymore -- they've all been changed to something else. Running the script again on the altered table will produce no errors, and will also produce no further change in the table, because the update statement has to match those old "ccode" values given in "ccchgtab.txt" and these are no longer in the table.

If you cannot rollback the table to get those old values again, then you must get a fresh dump of the affected records to create a new input file, and/or figure out how to do the updates without doing an exact match on the existing values of "ccode" -- something like:

update rtsdet set ccode='OFAA' where ccode like 'OFAA%' and ...

(if Informix supports the "%" expression in SQL string matches the way Oracle does, then you can include "%" as part of the "old" string pattern in your data file, and it will work as intended)

As for why the funny content got into the table in the first place, I don't see how that would result from either version of your code -- unless there was some subtlety that got munged or left out when you tried to post it.