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 $translator = SQL::Translator->new(
14 parser_args => { field_separator => "\t" },
19 Parses arbitrarily delimited text files. See the
20 Text::RecordParser manpage for arguments on how to parse the file
21 (e.g., C<field_separator>, C<record_separator>). Other arguments
30 Indicates that the columns should be scanned to determine data types
31 and field sizes. True by default.
35 A shortcut to sending filters to Text::RecordParser, will create
36 callbacks that trim leading and trailing spaces from fields and headers.
41 Field names will automatically be normalized by
42 C<SQL::Translator::Utils::normalize_name>.
46 # -------------------------------------------------------------------
49 use vars qw($VERSION @EXPORT);
53 use Text::ParseWords qw(quotewords);
54 use Text::RecordParser;
55 use SQL::Translator::Utils qw(debug normalize_name);
57 use base qw(Exporter);
61 # Passed a SQL::Translator instance and a string containing the data
64 my ( $tr, $data ) = @_;
65 my $args = $tr->parser_args;
66 my $parser = Text::RecordParser->new(
67 field_separator => $args->{'field_separator'} || ',',
68 record_separator => $args->{'record_separator'} || "\n",
70 header_filter => \&normalize_name,
73 $parser->field_filter( sub { $_ = shift || ''; s/^\s+|\s+$//g; $_ } )
74 unless defined $args->{'trim_fields'} && $args->{'trim_fields'} == 0;
76 my $schema = $tr->schema;
77 my $table = $schema->add_table( name => 'table1' );
80 # Get the field names from the first row.
83 my @field_names = $parser->field_list;
85 for ( my $i = 0; $i < @field_names; $i++ ) {
86 my $field = $table->add_field(
87 name => $field_names[$i],
92 is_auto_increment => undef,
93 ) or die $table->error;
96 $table->primary_key( $field->name );
97 $field->is_primary_key(1);
102 # If directed, look at every field's values to guess size and type.
105 defined $args->{'scan_fields'} &&
106 $args->{'scan_fields'} == 0
108 my %field_info = map { $_, {} } @field_names;
109 while ( my $rec = $parser->fetchrow_hashref ) {
110 for my $field ( @field_names ) {
111 my $data = defined $rec->{ $field } ? $rec->{ $field } : '';
112 my $size = [ length $data ];
115 if ( $data =~ /^-?\d+$/ ) {
119 $data =~ /^-?[,\d]+\.[\d+]?$/
121 $data =~ /^-?[,\d]+?\.\d+$/
127 map { s/,//g; length $_ || 1 } split( /\./, $data );
128 $size = [ $w + $d, $d ];
135 next unless defined $size->[ $i ];
136 my $fsize = $field_info{ $field }{'size'}[ $i ] || 0;
137 if ( $size->[ $i ] > $fsize ) {
138 $field_info{ $field }{'size'}[ $i ] = $size->[ $i ];
142 $field_info{ $field }{ $type }++;
146 for my $field ( keys %field_info ) {
147 my $size = $field_info{ $field }{'size'} || [ 1 ];
149 $field_info{ $field }{'char'} ? 'char' :
150 $field_info{ $field }{'float'} ? 'float' :
151 $field_info{ $field }{'integer'} ? 'integer' : 'char';
153 if ( $data_type eq 'char' && scalar @$size == 2 ) {
154 $size = [ $size->[0] + $size->[1] ];
157 my $field = $table->get_field( $field );
158 $field->size( $size );
159 $field->data_type( $data_type );
168 # -------------------------------------------------------------------
173 Darren Chamberlain E<lt>darren@cpan.orgE<gt>,
174 Ken Y. Clark E<lt>kclark@cpan.orgE<gt>.
178 Text::RecordParser, SQL::Translator.