most recent date

[Lovins, Brian (lovinsbk)]

Good afternoon list,

I have a large data set of inmates and need to determine the offense that is associated with the most recent release date.  The file is organized as a flat file with each case having the following variables:

Id Sdate.1 to sdate.10   chargedsc.1 to chargedsc.10

with sdate as sentence date and chargedsc as the description of the charge (string).  There may only be 1 sentence date or up to 10.  In addition some of the dates could be the same date.  If there is a sentence date there is always a charge description.  It is currently organized in the order of most serious offense so if there is a tie on dates the date closest to the 1st date should be fine.

Thanks
Brian

====================To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Maguin, Eugene]

Brian,

Right now you have a 'wide format' dataset. You can work the problem in that
format using a loop-end loop structure easily enough. But you can also work
the problem in a 'long format' dataset. This sentence: 'It is currently
organized in the order of most serious offense so if there is a tie on dates
the date closest to the 1st date should be fine.', is ambigious. I'll guess
that you mean that when there are multiple date variables with the same
value, the associated charges for that date set are ordered so that the most
serious charge is the first listed in date set. The other assumption is that
the sentence date is perfectly related to the releases date. Warning. Syntax
is not tested. This is where I'd start.

First, convert the dataset to a long format.

Varstocases make sdate from Sdate.1 to sdate.10/
   make chargedsc from chargedsc.1 to chargedsc.10.

Aggregate outfile=*/break=id sdate/chargedesc=first(chargedesc).

*  That reduces the dataset to one with exactly one record per date.
*  Now, aggregate to save the record for the last date.

Aggregate outfile=*/break=id/sdate chargedesc=last(sdate chargedesc).

Gene Maguin


>>I have a large data set of inmates and need to determine the offense that
is associated with the most recent release date.  The file is organized as a
flat file with each case having the following variables:

Id Sdate.1 to sdate.10   chargedsc.1 to chargedsc.10

with sdate as sentence date and chargedsc as the description of the charge
(string).  There may only be 1 sentence date or up to 10.  In addition some
of the dates could be the same date.  If there is a sentence date there is
always a charge description.  It is currently organized in the order of most
serious offense so if there is a tie on dates the date closest to the 1st
date should be fine.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD