|
midrange.com
Actually, soundex is supported directly in SQL. For performance reasons,
you'd want to create a "Search Key" file that hold the soundex values. I'd
probably add a few fields to describe the file, field, and ordinal value, so
that this one index can look into many files and fields.
Looking for John Smith? Is that JOHN or JON? SMITH or SMYTHE? Soundex or
metaphone renders the word to a phonetic hash value. In soundex, JOHN is
J500 (Metaphone: JN) and SMITH is S530 (SM0). I personally prefer metaphone
since the output is generally human readable, so I find it easier to work
with. I translated my Metaphone procedure from pascal (re-published in Joe
Celko's "SQL for Smarties") into RPG, then later into a procedure.
* METAPHONE-------------------------
D Metaphone PR 6a
D WordIn 128a Value
P Metaphone B
* Procedure Interface -----------------------------------
D*........................................................
D* Metaphone - An implementation adapted from source code
D* published in Joe Celko's "SQL for smarties". Source
D* translated to RPGIV from Pascal implementation written
D* by Terry Smithwick (1991).
D*
D* Usage: Metaphone( WordIn Char 128 ) returns: Char 6
D*
D* Eval $MetaCode = Metaphone($CityName)
D* {KLFLNT} {CLEVELAND}
D*........................................................
D Metaphone PI 6a
D WordIn 128a Value
* Local data --------------------------------------------
D Soundex S 6a
D i S 5i 0
D l S 5i 0
D n S 5i 0
D silent S n
D new S n
D last S 1a
D this S 1a
D next S 1a
D nnext S 1a
D m S 6a
D d S 2a
*
D VowelData DS
D 1a Inz('A')
D 1a Inz('E')
D 1a Inz('I')
D 1a Inz('O')
D 1a Inz('U')
D VowelSet 1a Dim(5) Overlay(VowelData)
*
D FrontVData DS
D 1a Inz('E')
D 1a Inz('I')
D 1a Inz('Y')
D FrontVSet 1a Dim(3) Overlay(FrontVData)
*
D VarSonData DS
Variable Sound
D 1a Inz('C')
D 1a Inz('S')
D 1a Inz('T')
D 1a Inz('G')
D VarSonSet 1a Dim(4) Overlay(VarSonData)
* Begin Metaphone Soundex Function ----------------------
C If WordIn = ''
C Return ' '
C EndIf
C* If first 2 letters in list of two letter consonnants, drop the leading
character.
C Eval d = %subst(WordIn:1:2)
C If d = 'KN' or
C d = 'GN' or
C d = 'PN' or
C d = 'PF' or
C d = 'AE' or
C d = 'WR'
C Eval WordIn = %subst(WordIn:2:%len(WordIn)-1)
C EndIf
C* If first letter is X, change it to S.
C If %subst(WordIn:1:1)='X'
C Eval %subst(WordIn:1:1)='S'
C EndIf
C* If first letters are WH, change it to W.
C If d = 'WH'
C Eval WordIn = 'W'+%subst(WordIn:3:%len(WordIn)-2)
C EndIf
C
C* Initialize values for looping
C Eval l = %len(%trim(WordIn))
C Eval m = ''
C Eval new = *On
C Eval n = 1
C
C DoW %len(%trim(m)) <= 6 and n <= l
C
C If n > 1
C Eval last = %subst(WordIn:n-1:1)
C Else
C Eval last = x'00'
C EndIf
C
C Eval this = %subst(WordIn:n:1)
C
C If n < l
C Eval next = %subst(WordIn:n+1:1)
C Else
C Eval next = x'00'
C EndIf
C
C If n+1 < l
C Eval nnext = %subst(WordIn:n+2:1)
C Else
C Eval nnext = x'00'
C EndIf
C
C If n > 1 and (this=last) and this<>'C'
C Eval n = n + 1
C Iter
C EndIf
C
C If new
C
C this Lookup VowelSet 10
C If *In10 and n = 1
C Eval m = this
C EndIf
C
C Select
C When this = 'B'
--------------B
C If not ((n=l) and last='M')
-MB is silent
C Eval m = %trim(m) + 'B'
C EndIf
C
C When this = 'C'
--------------C
C next Lookup FrontVSet 10
C If not ((last='S') and *In10)
-SC(E,I,Y) silent
C
C If next = 'I' and nnext = 'A'
-CIA- = X
C Eval m = %trim(m) + 'X'
C Else
C If *In10
C Eval m = %trim(m) + 'S'
-c(e,i,y) = S
C Else
C
C If next = 'H' and last = 'S'
-SCH
C Eval m = %trim(m) + 'K'
CH = K
C Else
C If next = 'H'
-CH-
C nnext Lookup VowelSet 10
C If n=1 and n+2<=l and not *In10
C Eval m = %trim(m) + 'K'
C Else
C Eval m = %trim(m) + 'X'
C EndIf
C Else
C Eval m = %trim(m) + 'K'
C = K
C EndIf
C EndIf
C EndIf
C EndIf
C EndIf
C
C When this = 'D'
--------------D
C nnext Lookup FrontVSet 10
C If next = 'G' and *In10
C Eval m = %trim(m) + 'J'
C Else
C Eval m = %trim(m) + 'T'
C EndIf
C
C When this = 'G'
--------------G
C nnext Lookup VowelSet 10
C Eval silent = (next='H') and *In10
-GHx
C If (n>1) and (((n+1)=l) or
C ((next='N') and (nnext='E') and
-GNED
C (%subst(WordIn:n+3:1)='D') and
C ((n+3)=l)) or (last='I') and
-IGN
C (next='N'))
C Eval silent = *On
C EndIf
C next Lookup FrontVSet 10
C If (n>1) and (last='D') and *In10
-DG(E,I,Y)
C Eval silent = *On
C EndIf
C
C If not silent
C next Lookup FrontVSet 10
C If *In10
C Eval m = %trim(m) + 'J'
C Else
C Eval m = %trim(m) + 'K'
C EndIf
C EndIf
C
C When this = 'H'
--------------H
C last Lookup VarSonSet 10
C next Lookup VowelSet 11
C If not ((n=l) or *In10) and *In11
C Eval m = %trim(m) + 'H'
C EndIf
C
C When this = 'F' or
--------------F
C this = 'J' or
J
C this = 'L' or
L
C this = 'M' or
M
C this = 'N' or
N
C this = 'R'
R
C Eval m = %trim(m) + this
C
C When this = 'K'
--------------K
C If last <> 'C'
C Eval m = %trim(m) + 'K'
C EndIf
C
C When this = 'P'
--------------P
C If next = 'H'
C Eval m = %trim(m) + 'F'
C Add 1 n
C Else
C Eval m = %trim(m) + 'P'
C EndIf
C
C When this = 'Q'
--------------Q
C Eval m = %trim(m) + 'K'
C
C When this = 'S'
--------------S
C If (next = 'H') or
C ((n>1) and (next='I') and
C (nnext='O' or nnext='A'))
C Eval m = %trim(m) + 'X'
C Else
C Eval m = %trim(m) + 'S'
C EndIf
C
C When this = 'T'
--------------T
C If (n=1) and (next='H') and (nnext='O')
THO-
C Eval m = %trim(m) + 'T'
C Else
C If (n>1) and (next='I') and
-TI(O,A)-
C (nnext='O' or nnext='A')
C Eval m = %trim(m) + 'X'
C Else
C If (next='H')
-TH- = 0
C Eval m = %trim(m) + '0'
C Else
C If not ((next='C') and (nnext='H'))
-TCH
C Eval m = %trim(m) + 'T'
C EndIf
C EndIf
C EndIf
C EndIf
C
C When this = 'V'
--------------V
C Eval m = %trim(m) + 'F'
C
C When this = 'W' or
--------------W
C this = 'Y'
Y
C next Lookup VowelSet 10
C If *In10
C Eval m = %trim(m) + this
C EndIf
C
C When this = 'X'
--------------X
C Eval m = %trim(m) + 'KS'
C
C When this = 'Z'
--------------X
C Eval m = %trim(m) + 'S'
C
C EndSL
C
C EndIf
C Eval n = n + 1
C EndDo
C Return m
P Metaphone E
Eric DeLong
Sally Beauty Company
MIS-Project Manager (BSG)
940-898-7863 or ext. 1863
-----Original Message-----
From: Booth Martin [mailto:booth@xxxxxxxxxxxx]
Sent: Friday, September 10, 2004 12:46 PM
To: Midrange Systems Technical Discussion
Subject: RE: fast search or scan for words in selected field
Well, how about this scenario then? Would this be a successful alternative?
Could a file of the search words be made, as is presently being done? Then
use SQL to search that base? An advantage of SQl then would be to allow
searches on matching two or more of the keywords?
But the discussions on Soundex, etc. seems like a real winner though. Where
would we learn more about that?
---------------------------------
Booth Martin
http://www.martinvt.com
---------------------------------
-------Original Message-------
From: Midrange Systems Technical Discussion
Date: 09/10/04 12:34:46
To: Midrange Systems Technical Discussion
Subject: RE: fast search or scan for words in selected field
I'm with Joe on this one. I can't see how indices would help with
substring searches. Then again, just because I can't comprehend it
doesn't mean it wouldn't.
Rob Berendt
--
Group Dekko Services, LLC
Dept 01.073
PO Box 2000
Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.
As an Amazon Associate we earn from qualifying purchases.
This mailing list archive is Copyright 1997-2025 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.
