My understanding of the original problem posed by Paul Gauthier on Friday was to do "transaction matching" using a "tolerant sort". My assumption is that records arrive in a database; they contain what SHOULD be the same value but they may have been computed by completely different systems using different algorithms, different hardware, etc. We need to do a tolerant sort which group these records together using fuzzy matching of the floating-point values, but preserves the order that they have in the database within the groups.
If this is so, then any algorithm which uses the first value or the lowest value, or "any SIMPLE strategy" as the basis for deciding whether other records are tolerantly equal, runs the (theoretical) risk of excluding values that are within the tolerance of some theoretical "correct" value. In practice, I suspect that this problem does not actually occur because the variation in values is so small that it does not matter which value you pick as the basis for these comparisons. -----Original Message----- From: Raul Miller [mailto:[EMAIL PROTECTED] Sent: 6. juni 2007 20:34 To: General forum Subject: Re: [Jgeneral] Re: Challenge to expert J'ers On 6/6/07, Morten Kromberg <[EMAIL PROTECTED]> wrote: > I may be missing something, but I think this depends on which of X1, X2 and > X3 are "the correct value". If X2 is the quote correct unquote answer and X1 > and X3 are the result of using slightly different computations to arrive at > it, then the application is going to want to see X1, X2 and X3 all end up in > the same bucket when the transactions are sorted. If X1 is correct then > maybe X3 should be left out if it is not tolerantly equal. I do not understand what you are getting at, here. > I think this means that arriving at the "desired" result requires > information which a general "tolerant grade up" cannot have (which of X1, X2 > and X3 are "closest to the right result"). To my knowledge, tolerant i.~ uses "intolerant grade up". Tolerant grade up is a completely different issue, as far as I can tell. -- Raul ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
