Date:         Wed, 22 Feb 2006 11:06:50 -0800
Reply-To:     "Choate, Paul@DDS" <pchoate@DDS.CA.GOV>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Choate, Paul@DDS" <pchoate@DDS.CA.GOV>
Subject:      Re: studying repeats
Comments: To: "lrbowes2@yahoo.com" <lrbowes2@YAHOO.COM>
Content-Type: text/plain; charset="us-ascii"

If the data aren't too large (<64k lines) I often put this sort of data
into Excel (with the libname engine) and use a pivot table to look at
the problems.  For me it's very interactive and lets me rapidly drill
across and down and inspect the situation.  I then use this to inform my
SAS programming. If the data are large I often subset or sample it and
do the same.

I agree that keeping your key fields separate is important - using
multiple level sorts instead.  This allows one to also look for problems
within the levels independently.

hth

Paul Choate
DDS Data Extraction
(916) 654-2160

> -----Original Message-----
> From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
> lrbowes2@yahoo.com
> Sent: Tuesday, February 21, 2006 2:02 PM
> To: SAS-L@LISTSERV.UGA.EDU
> Subject: studying repeats
>
> Hi.
>
> I need to look at how many times and which various combinations of
data
> are repeated (zip code and time period).  (Ideally, there should only
> be one instance of each zip code, time period combo, so I want to
learn
> more about cases where that is false.)  I created a variable that is
> the concatenation of the values of zip code and time period.  What is
> the best way of getting summary stats on the repeats?
>
> I don't want to just use nodup because I want to know what was
repeated
> and how many times it was repeated.
>
> Ideally, I guess I would maybe do a proc freq on the concatenated
> variable, but I want to get rid of all instances where the frequency
is
> 1.
>
> Thanks.
>
> -Lori