Re: QSORT question

On 12/19/06, Jon Paris <Jon.Paris@xxxxxxxxxxxxxx> wrote:

Since multiplication is a pretty slow thing to do - wouldn't it be far
simpler - and _way_ more obvious - to code the sort this way?  This is how I
always do multi-key sorts, which is what this really is.

      select;
          when E1.Year < E2.Year;
               return -1;
          when E1.Year > E2.Year;
               return 1;
          other;
          select;
               when E1.Month < E2.Month;
                    return -1;
               when E1.Month > E2.Month;
                    return 1;
               other;
                    return 0;
          endsl;
      endsl;


Jon Paris
Partner400

www.Partner400.com

Which would get deeply nested for a bunch of keys. I have a generic
file comparison report, which needs to be able to compare 1 to n key
fields; I do it like:

<PRE><CODE>
p compKey         b
d                 pi            10i 0
/free
 $I = 0;
 DoU ($I = keyCnt Or memcmpstr <> 0);
    $I = $I + 1;
    memcmpstr = memcmp( @bfr1@($I)
                      : @bfr2@($I)
                      : @fldLnB($I)
                      );
 EndDo;

 Return memcmpstr;
/end-free
<CODE></PRE>
Where:

There are $I key fields

@bfr1@ is an array of pointers to the first character of the keyfields in the
first file to be compared.

@bfr2@ is an array of pointers to the first character of the keyfields in the
second file to be compared.

@fldLnB is an array containing the length of the keyfields.


This also has the added advantage of allowing me to determine how many
keys at run time, and not compile time.

www.brilligware.com/cp1030.html

Chris
--
chris@xxxxxxxxx           | You know you've achieved perfection in
www.pando.org             | design, not when you have nothing more
                         | to add, but when you have nothing more
                         | to take away. Antoine de Saint-Exupery