Jeff Thies wrote:
> Anyone have any thoughts on subclassing AnyData for virtmaps?
> Virtmaps have two fields delimited by white space.
You can just use AnyDatat's CSV format with space as a separator and
trim set to 1 to eliminate the extra spaces:
$dbh->func('test','CSV',[<DATA>],{field_sep=>' ',trim=>1},'ad_import');
> I would think that for reading (SELECT) them you could pass in:
> field_sep,\s on the flags;
>
> And for writing (INSERT):
> field_sep,\t on the flags;
>
> I'm completely confused on how you would do an UPDATE.
Well, subclassing AnyData is alot simpler than you imagine. All you
need to supply is a read_field and write_field and AnyData will handle
things like deleting, updating etc. Here's an entire subclass that
would handle all the same SQL as the other formats for what you describe:
package AnyData::Format::Vertmap
use AnyData::Format::Base;
use vars qw( @ISA $VERSION);
@AnyData::Format::Vertmap::ISA = qw( AnyData::Format::Base );
sub new {
my $class = shift;
my $self = shift || {};
return bless $self, $class;
}
sub read_fields {
my $self = shift;
my $str = shift || return undef;
my @fields = split /\s+/,$str;
}
sub write_fields {
return join "\t",@_;
}
1;
That's all, stick in in your AnyData/Format directory and you can use it
the same as Pipe, CSV, etc.
--
Jeff
