#!/usr/bin/perl -s
use strict;
use Text::TabDB;
use utf8::all; 

our($fs,$rm ,$o);
my $coms = not $rm;
my $cond = shift or die("usage: tabselect [-fs=#] condition file*\n");

Text::TabDB::tabselect({fs=>$fs,coms=>$coms,($o?(o=>$o):())},$cond,@ARGV);

__END__

=head1 NAME

tabselect - textual relational databases select

=head1 SYNOPSIS

 tabselect [-fs=#] condition tab* > f
 iconv -f latin -t utf8 tab1 | tabselect '$F1 eq $F4'


=head1 DESCRIPTION

condition is a perl expression that may have

 $id    variable with field whose idenfier is "id"
 $_2    variable with field 2

input should be Unicode UTF8.

If no tab are provided stdin is used.

=head1 FORMATS

each input table: 

  country:capital       first line - name of the fields
  Portugal:Lisboa       tuple 1
  France:Paris          tuple 2
  Spain:Madrid          tuple n

=head1 Conditions syntax

Condition follows the Perl syntax for conditions; C<$_1> is the first field;
C<$id> is the field corresponding to first line C<id>

== Example

  tabselect '$_2' f.tab        capital is defined
  tabselect '$capital' f.tab   the same
  tabselect 'not $_2' f.tab    capital is empty
  tabselect '$_1 =~ /^P/' f.tab   Country starts with 'P'
  
=head1  Options

  -fs='!'  changes the field separator (default calculated 
              from schema (first line) or ":")

  -o=file  output file
  -rm      remove comments (#...), directives (%...) and emptylines
             (def: keep them untouched)

=head1 AUTHOR

J.Joao Almeida, jj@di.uminho.pt

=head1 SEE ALSO

perl(1).

=cut