Date:         Thu, 7 Jul 2005 15:00:22 -0400
Reply-To:     "Howard Schreier <hs AT dc-sug DOT org>" <nospam@HOWLES.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Howard Schreier <hs AT dc-sug DOT org>" <nospam@HOWLES.COM>
Subject:      Re: Identical inputs different results!

It's hard to say much without being able to see specifics.

In my experience, observations which are truly identical give the same
results. I mean identical byte by byte; you may need hexadecimal display to
confirm that.

To avoid divergent results from data which are not quite identical, look at
the ROUND function or the FUZZ option. Again, it's hard to say what's
applicable or appropriate without having some details in hand.

On Thu, 7 Jul 2005 07:14:28 -0400, Dennis Diskin <ddiskin@GMAIL.COM> wrote:

>Aaron.
> Please show examples of identical data giving different results,
including
>code.
> Regards,
>Dennis Diskin
>
> On 7/7/05, aaronfude@gmail.com <aaronfude@gmail.com> wrote:
>>
>> Hi,
>>
>> My database has many observations (lines), but some observations are
>> pairwise identical.
>>
>> I do identical computations on all observations and then sort the
>> results. If the results are identical, then get identical indices in
>> the sort. However, every once in a while (about 1 time out 20) the
>> identical observations get different indices, presumably because they
>> got rounded in different ways.
>>
>> Is there a way to avoid this?
>>
>> Many thanks in advance!
>>
>> Aaron Fude
>>