A 'LOGICAL' SORT

	These tested patterns provide a method to order entries in a data-
	base or word processing environment in a desired manner.  'Well, 
	isn't that what a database is for?', you ask.

	Yes and No.  Where material is desired to have seperate sections
	within a database itself this relatively simple method takes the
	pain out of a lot of programming that would usually be employed to
	do the job.

	Are there other ways? - Sure, I guess.

	How does it work?

	Using the sequences below segregates a database into two basic areas 
	based on including them as the first or 'key' field in a prescribed
	sort.  The first 'section' (entries that will appear at the top) is
	composed of records in which NO entry (a 'null') has been made in 
	the 'key' field.

	This is also the prime area where real sorting will occur. As the 
	'key' fields are all identical, sorting defaults to the next chosen
	field - the real data.
	
	The second 'section' encounters the sequences in the 'key' field and
	organizes around them.  This area could be used for things such as a
	Table of Contents etc. ...  As currently designed this will provide
	up to 1073 reserved records using one key field of two text 
	characters in length.

	note: by using the same seguence of codes in a second 'field' these 
	can be resued thus extending the range of reserved area.
	e.g. by setting the first or 'key' field to '0' and using the 1073 
	combinations in the second field and then changing the key field to 
	"00" (i.e. following the same pattern below) you can reuse the 
	entire sequence again in the second field for another 1073 entries. 
	Changing the key field to '01' gets another 1073 entries and so on 
	for 1073 total passes.  This produces over one million possible 
	entries using only two fields of two text characters each (or one
	four character field ; just don't forget the 'space' after any 
	'single' leading code; e.g. "0 z").



				The Sorting Sequences


	Alpha/Numeric (these will appear above alphabetical only)

	"0" then 
	"00"-"09" then 
	"0a"-"0z" (total 37)
	--
	"1" then
	"10"-"19" then
	"1a" - "1z" (total 37)
	--
	"2" then
	"20"-"29" then
	"2a"-"2z" (total 37); use this section for Level 1 World Menu area

	....and so forth for 3-9z; i.e. 10 sets of 37 = 370 reserved records.
	(if we limit it to 0a - 9z (26 times 10) = 260 reserved records)

	--

	Alphabetical only:


	"a" then
	"aa"-"az"   (total 27)  then
	"b" then
	"ba"-"bz"   (total 27) ...

	Following the same pattern: next letter 'c' followed by next ca-cz
	through end of alphabet gets 26x27= 702 alfa only reserved records.

	If we limit to dual alpha only (aa-zz) we get 26 sets of 26 pairs or 
	676 reserved records.

	Further : limited sets from each type = 260 + 676 = 913 records
	with a total of 913x913= a big bunch if codes are used in two fields
	(which can be combined into one field if so desired; it just seems
	easier at the planning stage to think in descrete limited terms)

	Using the non-limited: 702+371=1073  or 1073x1073 using two fields.

	Another variation: use a limited triple-alpha sequence in a 
	three character field (e.g. aaa-zzz) = 26x26x26 = 17,576 records

	a
	aa
	aaa - aaz (28 records)
	aba - azz (25x26)

	b - zzz (25x26x26)

	HOPEFULLY ONE OF THESE IS ENOUGH. 
	(if not: add another field or more width and due the math.)


	One final note: data transfered to a word processor sorts the same
	when the sort is limited to each area seperately. 

	i.e. where a <null> (no entry) is used as opposed to the codes above.
	Keep them seperate in a word processor.


	The sort sequence:
	Key Field1 + Key Field2 (if used) + any chosen field(s) for the 
	primary sort area.


	A.J. Ward 
	Dec. 20, 2000



	(c) 2000