1 package SQL::Translator::Parser::xSV;
5 SQL::Translator::Parser::xSV - parser for arbitrarily delimited text files
10 use SQL::Translator::Parser::xSV;
12 my $translator = SQL::Translator->new(
14 parser_args => { field_separator => "\t" },
19 Parses arbitrarily delimited text files. See the
20 Text::RecordParser manpage for arguments on how to parse the file
21 (e.g., C<field_separator>, C<record_separator>). Other arguments
30 Indicates that the columns should be scanned to determine data types
31 and field sizes. True by default.
35 A shortcut to sending filters to Text::RecordParser, will create
36 callbacks that trim leading and trailing spaces from fields and headers.
41 Field names will automatically be normalized by
42 C<SQL::Translator::Utils::normalize_name>.
47 use vars qw($VERSION @EXPORT);
51 use Text::ParseWords qw(quotewords);
52 use Text::RecordParser;
53 use SQL::Translator::Utils qw(debug normalize_name);
55 use base qw(Exporter);
59 # Passed a SQL::Translator instance and a string containing the data
62 my ( $tr, $data ) = @_;
63 my $args = $tr->parser_args;
64 my $parser = Text::RecordParser->new(
65 field_separator => $args->{'field_separator'} || ',',
66 record_separator => $args->{'record_separator'} || "\n",
68 header_filter => \&normalize_name,
71 $parser->field_filter( sub { $_ = shift || ''; s/^\s+|\s+$//g; $_ } )
72 unless defined $args->{'trim_fields'} && $args->{'trim_fields'} == 0;
74 my $schema = $tr->schema;
75 my $table = $schema->add_table( name => 'table1' );
78 # Get the field names from the first row.
81 my @field_names = $parser->field_list;
83 for ( my $i = 0; $i < @field_names; $i++ ) {
84 my $field = $table->add_field(
85 name => $field_names[$i],
90 is_auto_increment => undef,
91 ) or die $table->error;
94 $table->primary_key( $field->name );
95 $field->is_primary_key(1);
100 # If directed, look at every field's values to guess size and type.
103 defined $args->{'scan_fields'} &&
104 $args->{'scan_fields'} == 0
106 my %field_info = map { $_, {} } @field_names;
107 while ( my $rec = $parser->fetchrow_hashref ) {
108 for my $field ( @field_names ) {
109 my $data = defined $rec->{ $field } ? $rec->{ $field } : '';
110 my $size = [ length $data ];
113 if ( $data =~ /^-?\d+$/ ) {
117 $data =~ /^-?[,\d]+\.[\d+]?$/
119 $data =~ /^-?[,\d]+?\.\d+$/
125 map { s/,//g; length $_ || 1 } split( /\./, $data );
126 $size = [ $w + $d, $d ];
133 next unless defined $size->[ $i ];
134 my $fsize = $field_info{ $field }{'size'}[ $i ] || 0;
135 if ( $size->[ $i ] > $fsize ) {
136 $field_info{ $field }{'size'}[ $i ] = $size->[ $i ];
140 $field_info{ $field }{ $type }++;
144 for my $field ( keys %field_info ) {
145 my $size = $field_info{ $field }{'size'} || [ 1 ];
147 $field_info{ $field }{'char'} ? 'char' :
148 $field_info{ $field }{'float'} ? 'float' :
149 $field_info{ $field }{'integer'} ? 'integer' : 'char';
151 if ( $data_type eq 'char' && scalar @$size == 2 ) {
152 $size = [ $size->[0] + $size->[1] ];
155 my $field = $table->get_field( $field );
156 $field->size( $size );
157 $field->data_type( $data_type );
170 Darren Chamberlain E<lt>darren@cpan.orgE<gt>,
171 Ken Y. Clark E<lt>kclark@cpan.orgE<gt>.
175 Text::RecordParser, SQL::Translator.