Date: Mon, 17 Apr 2000 09:58:14 -0500
Reply-To: Tim Dunsworth <Tim.Dunsworth@METROSTATE.EDU>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Tim Dunsworth <Tim.Dunsworth@METROSTATE.EDU>
Subject: Re: syntax question
Content-Type: text/plain; charset=US-ASCII
We may need a little further detail to be sure what to suggest here. If it is the original erroneous record that is flagged with negative values, then Hector's suggestion is a good one and should work for you. As I read your description, though, it sounded like it is the NEW record that is a mixture of corrected values and negative flag values. If so, you certainly want those corrected values, and your processing task could get a little more complicated.
If the new record has corrections where needed, plus copies of most other fields and just a few specific flag fields that are negative, then you could sort on ID and a flag field and use the aggregate function, breaking on id and picking the First value for most fields (to get the corrected values) and the Last value for the flag fields (to get their original non-negative values). But if things are not that systematic, with only corrected fields having non-negative values for example, then I am not sure how this can easily be done.
Please clarify exactly what the situation is, then if the worst case applies, maybe someone else has a suggestion for how to deal with it.
>>> Scott Compton <scompton@ACPUB.DUKE.EDU> 04/17 7:26 AM >>>
I'm working with an insurance claims data file in which data entry
errors are handled by adding a new row of data and then making key
values negative in that row, similar to when you return something
purchased with a credit card and it then shows up on the monthly
statement as a negative transaction.
My question, is there a way to get SPSS to delete only those rows that
have negative values on specific key variables?
Thanks in advance for any help with this problem.