[thelist] importing a non-standard datadump
Luther, Ron
ron.luther at hp.com
Wed Feb 23 07:49:45 CST 2005
Brian Cummiskey noted:
>>Thanks for the help guys. After doing search and replaces for most of
>>the day yesterday, i thought i had a decent file to play with.
>>go to import it-
>>And now, i am missing a field (most likely a ,"", field). Of course,
in
>>24,000 records, and tons of 2nd address lines filled in this way,
looks
>>like its time to go back to the client and have them send me another
>>dump, and start all over :(
Hi Brian,
Sorry. I guess we should have spelled out 'Rule 1' more explicitly:
"Always work from a backup of the raw client data."
;-(
Here is another idea that might help some: Take a good hard look at
your raw data. It may be possible to determine how many good lines
you 'should' have when you finally get things cleaned up. {Maybe
something as easy as a count of all 'pipes' divided by two. Maybe
something a bit trickier.} The value here is that it would give you
a 'sanity cross check' to improve your confidence that you haven't
misplaced any data records once you finally think you might be
done.
BTW - This data really isn't that 'weird' ... I don't think it's that
uncommon to have to parse 'print image' mainframe reports into local
data tables ... it can get a whole lot uglier than what it looks like
you are playing with here.
Good Luck!!
RonL.
More information about the thelist
mailing list