package SQL::Translator::Parser::xSV;
# -------------------------------------------------------------------
-# $Id: xSV.pm,v 1.6 2003-05-07 20:36:59 kycl4rk Exp $
-# -------------------------------------------------------------------
-# Copyright (C) 2003 Ken Y. Clark <kclark@cpan.org>,
-# darren chamberlain <darren@cpan.org>
+# Copyright (C) 2002-2009 SQLFairy Authors
#
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License as
(e.g., C<field_separator>, C<record_separator>). Other arguments
include:
+=head1 OPTIONS
+
=over
=item * scan_fields
Indicates that the columns should be scanned to determine data types
-and field sizes.
+and field sizes. True by default.
=item * trim_fields
A shortcut to sending filters to Text::RecordParser, will create
callbacks that trim leading and trailing spaces from fields and headers.
+True by default.
=back
Field names will automatically be normalized by
-C<SQL::Translator::Utils::normalize>.
+C<SQL::Translator::Utils::normalize_name>.
=cut
use strict;
use vars qw($VERSION @EXPORT);
-$VERSION = sprintf "%d.%02d", q$Revision: 1.6 $ =~ /(\d+)\.(\d+)/;
+$VERSION = '1.60';
use Exporter;
use Text::ParseWords qw(quotewords);
# Passed a SQL::Translator instance and a string containing the data
#
sub parse {
- my ($tr, $data) = @_;
-
+ my ( $tr, $data ) = @_;
my $args = $tr->parser_args;
my $parser = Text::RecordParser->new(
field_separator => $args->{'field_separator'} || ',',
header_filter => \&normalize_name,
);
- $parser->field_filter( sub { $_ = shift; s/^\s+|\s+$//g; $_ } )
- if $args->{'trim_fields'};
+ $parser->field_filter( sub { $_ = shift || ''; s/^\s+|\s+$//g; $_ } )
+ unless defined $args->{'trim_fields'} && $args->{'trim_fields'} == 0;
- #
- # Create skeleton structure, mostly empty.
- #
- my $parsed = {
- table1 => {
- type => undef,
- indices => [ { } ],
- fields => { },
- },
- };
+ my $schema = $tr->schema;
+ my $table = $schema->add_table( name => 'table1' );
#
# Get the field names from the first row.
my @field_names = $parser->field_list;
for ( my $i = 0; $i < @field_names; $i++ ) {
- $parsed->{'table1'}{'fields'}{ $field_names[$i] } = {
- type => 'field',
- order => $i,
- name => $field_names[$i],
-
- # Default datatype is "char"
- data_type => "char",
-
- # default size is 8bits; something more reasonable?
- size => [ 255 ],
- null => 1,
- default => '',
- is_auto_inc => undef,
-
- # field field is the primary key
- is_primary_key => ($i == 0) ? 1 : undef,
+ my $field = $table->add_field(
+ name => $field_names[$i],
+ data_type => 'char',
+ default_value => '',
+ size => 255,
+ is_nullable => 1,
+ is_auto_increment => undef,
+ ) or die $table->error;
+
+ if ( $i == 0 ) {
+ $table->primary_key( $field->name );
+ $field->is_primary_key(1);
}
}
#
# If directed, look at every field's values to guess size and type.
#
- if ( $args->{'scan_fields'} ) {
+ unless (
+ defined $args->{'scan_fields'} &&
+ $args->{'scan_fields'} == 0
+ ) {
my %field_info = map { $_, {} } @field_names;
while ( my $rec = $parser->fetchrow_hashref ) {
for my $field ( @field_names ) {
my $data = defined $rec->{ $field } ? $rec->{ $field } : '';
- my $size = length $data;
+ my $size = [ length $data ];
my $type;
if ( $data =~ /^-?\d+$/ ) {
$type = 'integer';
}
- elsif ( $data =~ /^[\d.-]+$/ ) {
+ elsif (
+ $data =~ /^-?[,\d]+\.[\d+]?$/
+ ||
+ $data =~ /^-?[,\d]+?\.\d+$/
+ ||
+ $data =~ /^-?\.\d+$/
+ ) {
$type = 'float';
+ my ( $w, $d ) =
+ map { s/,//g; length $_ || 1 } split( /\./, $data );
+ $size = [ $w + $d, $d ];
}
else {
$type = 'char';
}
- if ( $size > $field_info{ $field }{'size'} ) {
- $field_info{ $field }{'size'} = $size;
+ for my $i ( 0, 1 ) {
+ next unless defined $size->[ $i ];
+ my $fsize = $field_info{ $field }{'size'}[ $i ] || 0;
+ if ( $size->[ $i ] > $fsize ) {
+ $field_info{ $field }{'size'}[ $i ] = $size->[ $i ];
+ }
}
$field_info{ $field }{ $type }++;
}
for my $field ( keys %field_info ) {
- $parsed->{'table1'}{'fields'}{ $field }{'size'} =
- [ $field_info{ $field }{'size'} ];
+ my $size = $field_info{ $field }{'size'} || [ 1 ];
+ my $data_type =
+ $field_info{ $field }{'char'} ? 'char' :
+ $field_info{ $field }{'float'} ? 'float' :
+ $field_info{ $field }{'integer'} ? 'integer' : 'char';
+
+ if ( $data_type eq 'char' && scalar @$size == 2 ) {
+ $size = [ $size->[0] + $size->[1] ];
+ }
- $parsed->{'table1'}{'fields'}{ $field }{'data_type'} =
- $field_info{ $field }{'char'} ? 'char' :
- $field_info{ $field }{'float'} ? 'float' : 'integer';
+ my $field = $table->get_field( $field );
+ $field->size( $size );
+ $field->data_type( $data_type );
}
}
- #
- # Field 0 is primary key, by default, so add an index
- #
- for ( $parsed->{'table1'}->{'indices'}->[0] ) {
- $_->{'type'} = 'primary_key';
- $_->{'name'} = undef;
- $_->{'fields'} = [ $field_names[0] ];
- }
-
- return $parsed;
+ return 1;
}
1;
# -------------------------------------------------------------------
=pod
-=head1 AUTHOR
+=head1 AUTHORS
Darren Chamberlain E<lt>darren@cpan.orgE<gt>,
-Ken Y. Clark E<lt>kclark@cpan.orgE<gt>.
+Ken Youens-Clark E<lt>kclark@cpan.orgE<gt>.
=head1 SEE ALSO
-Text::RecordParser.
+Text::RecordParser, SQL::Translator.
=cut