=pod

Code submission from: Paul Schaap (schaappa@ozemail.com.au)

fastFind 1.0.0 Release Notes

WHY
===
A very,very fast way to do a search on a very large file. Achieved by 
presorting the file and making the length of each field the same. You 
can then use the awesomely quick sysseek and sysread Perl functions.

INSTALLATION
============
ï Install MacPerl 5.1.4r4 or greater.(Thanx Matthias) And add Chris 
Nandors Sort Perl Module on the Mac, or, use Alpha or Word to sort the 
Index file.

USAGE
=====
1) Drop a file with the structure :-
  1. Name
  2. Address
  3. Suburb
  4. Postal Code
  5. Phone number
delimited by pipe "|" onto 'generateIndex Droplet' (e.g. generateIndex 
Input).

2) Open and run fastFind.pl then maybe use it as a require in an 
application or cgi.

CHANGES
=======
V1.0.0

ï Well none really, first cut.

CHEERS
schaapp@ozemail.com.au - I Only read email once a fortnight, sorry.




=cut

#!perl
######################################################################
##########
# PROGRAM : fastFind.pl
# BY      : Paul Schaap
# DATED   : 05/01/1998
# PURPOSE : Read Large files instantly
######################################################################
##########

# USAGE &fastFind(Index filename,Indexcount filename,key);
print &fastFind("indexFile","indexCount","056736426");

# FASTFIND ROUTINE
#==================
sub fastFind {
  $table = $_[0];
  $count = $_[1];

  # Read Number of Rows and Table structure
  open(PC,$count) or print "Failed to open $count\n";
  @indexInfo = split(/,/,<PC>);
  $indexRows = $indexInfo[0];
  $keyLength = $indexInfo[1];
  $value     = sprintf("%$keyLength.$keyLength"."s",$_[2]);
  $rowLength = $indexInfo[1];
  for($i=2;$i< @indexInfo;$i++){
    $rowLength       = $rowLength + $indexInfo[$i];
    $fieldLength[$i] = $indexInfo[$i];
  }
  close PC;
  $rowLength = $rowLength + 1;
  #print "$rowLength\n";

  # Go for it
  open(PH,$table) or return "Cannot open $table\n";
  $result = "";
  $low    = 0;
  $high   = $indexRows - 1;
  while($low <= $high){
    $mid = int (($low + $high) / 2);
    $off = $mid * $rowLength;
    #print "$mid $off\n";
    sysseek PH,$off,0;
    sysread PH,$rec,($rowLength-1);
    $key    = substr($rec,0,$keyLength);
    $result = substr($rec,$keyLength,$rowLength-$keyLength);
    #print "-$key-$rec-\n\n";
    if($key eq $value){
      close PH;
      return $result;
    } else {
      if($key gt $value){
        $high = $mid - 1;
      } else {
        $low = $mid + 1;
      }
    }
  }
  close PH;
  return "NOT FOUND";
}
1;
# END
#####################################################################