Date:         Thu, 6 Jan 2011 17:49:21 -0600
Reply-To:     Joe Matise <snoopy369@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Joe Matise <snoopy369@GMAIL.COM>
Subject:      Re: help with matching without replacement
Comments: To: Tracy Li <lisiqi77@yahoo.com>
In-Reply-To:  <201101062332.p06KicNs029630@wasabi.cc.uga.edu>
Content-Type: text/plain; charset=ISO-8859-1

How do you prevent id_match 3 from being in the sample twice?  Does it
matter which id the single id_match comes from?

I probably would suggest a simple data step BY ID YEAR SIZE_DIF, taking
FIRST.YEAR, and using a hash table (or array, if unique # of id_match is
small) to mark each id_match as used already.

-Joe

On Thu, Jan 6, 2011 at 5:32 PM, Tracy Li <lisiqi77@yahoo.com> wrote:

> Hi,
>
> I have a question regarding how to select a matched sample without
> replacement.
>
> Suppose my data looks like the following:
>
> id year id_match size_dif
> a 2000 1          12
> a 2000 2          22
> a 2000 3           1
> a 2000 4          23
> a 2001 2          24
> a 2001 5           2
> a 2001 6           4
> b 2000 3           5
> b 2000 4           7
> b 2000 7          32
>
> My goal is to find one and only one 'id-match' with the smallest possible
> size difference 'size-dif' with 'id' for each id-year.  And there should
> be no duplicates in the matched sample (that is, I don't want id_match=3
> in 2000 appear twice in the matched sample).
>
> Any ideas would be greatly appreciated!
>
> Tracy
>