See the table_info() method in DBI.

Cake walk.

Check out your SqlServer docs (or the online ones) for the system stored procedures sp_tables and sp_columns if the DBI module can't help you. Be sure to actually call those procs from DBI, though! Otherwise my perlmonk comment is wasted. ;)

Hanlon's Razor - "Never attribute to malice that which can be adequately explained by stupidity"

sub GetDatabases
{
    my @tempdbname;
    my @databases2;
    my @databases = DBI->data_sources("mysql");

    for my $x(0..$#databases)
    {
        @tempdbname = split(/:/,$databases[$x]);
        push @databases2, $tempdbname[$#tempdbname];
    }
    return @databases2;
}


sub GetTables
{
    my $database = shift;
    my @tables2;
    my $datasource = "dbi:mysql:dbname=$database;host=localhost;";

    my $dbh = DBI->connect($datasource,'dbuser','password' ,{PrintErro
+r => 0}) or return @tables2;

    my $statement = "SHOW TABLES";
    my $sth = $dbh->prepare($statement);
    my $rc = $sth->execute();
    if($rc ne "0E0")
    {
        my $tables = $sth->fetchall_arrayref;
        for my $xx(0..$#{$tables})
        {
            push @tables2,$tables->[$xx][0];
        }
    }

    return @tables2;
}

sub GetFields
{
    my $database = shift;
    my $table = shift;
    my @fields;
    my $counter = 0;
    my $test;

    my $datasource = "dbi:mysql:dbname=$database;host=localhost;";
    my $dbh = DBI->connect($datasource,'dbuser','password');
    my $statement = "DESCRIBE $table;";
    my $sth = $dbh->prepare($statement99);
    my $rc = $sth->execute();

    while($test = $sth->fetchrow_hashref)
    {
        $counter++;
        for my $t(keys %{$test})
        {
            #            field        field info
            $fields[$counter]{$t} = $test->{$t};
        }
    }
    return @fields;
}
[download]

use strict;
use DBI;
my $dbh = DBI->connect('DBI:database_type:database_name','user','passw
+ord') || die DBI->errstr;
my $sth = $dbh->prepare("SELECT * FROM table_name WHERE 0=1") || die $
+dbh->errstr;
$sth->execute() || die $sth->errstr;
my @column_names = @{$sth->{NAME}};
$sth->finish();
$dbh->disconnect();
[download]

While you should be commended to point at a lowest common denominator solution it should be pointed out that the LIMIT keyword is non-standard (and your code snippet is also missing the $sth->execute call :-).

It shouldn't be necessary to even fetch a single row for this. A WHERE clause that evaluates to false achieves this with minimal impact on the server:

use strict;
use DBI;
my $dbh = DBI->connect('DBI:database_type:database_name','user','passw
+ord') || die DBI->errstr;
my $sth = $dbh->prepare("SELECT * FROM table_name WHERE 0=1") || die $
+dbh->errstr;
$sth->excecute || die $sth->errstr;
my @column_names = @{$sth->{NAME}};
$sth->finish();  # may be unnecessary
$dbh->disconnect();
[download]

Update Fixed typo in WHERE clause...

Michael

The snippet doesn't work I can't print anything thing and WHERE 0=1 1 is not ODBC friendly.

In general your code is a good way to approach getting column names, but you get a bit tripped up on some of the specifics at least if the code has any hope of being portable.

my $sth=$dbh->prepare("SELECT * FROM table_name LIMIT 1")...
my @column_names = @{$sth->{NAME}};
[download]

1. LIMIT is not standard SQL and doesn't exist for many RDBMSs, unfortunately even "WHERE 1=0" isn't usable everywhere

2. $sth->{NAME} is not necessarily valid until after an execute.

3. It's best to specify $sth->{NAME_lc} or $sth->{NAME_uc} since different databases store the column names differently even if they ignore the case in actual SQL.

And your Perl question is?

Most RDBMSs I've seen have built-in capability to list tables, and to list columns of a table, yes. Exacly how that's done varies.