Most databases have system tables that contain that information. You'll have to consult your documentation. However, to find just the fieldnames from a table, this hack should do it (though I probably wouldn't use something like this):
my $dbh = DBI->connect( $connection,$user, $pass,{RaiseError => 1});
my $sql = 'SELECT TOP 1 * FROM someTable';
my $sth = $dbh->prepare( $sql );
$sth->execute;
my $fields = $sth->fetchrow_hashref;
$sth->finish;
my @fields = keys %$fields;
Cheers,
Ovid
Join the Perlmonks Setiathome Group or just click on the the link and check out our stats. | [reply]
[d/l] |
Yep, Ovid is dead on here -- the information for which you are looking is stored in the Oracle data dictionary.
I hope the following example helps. (Note: There are more fields in dba_tab_columns... you might want to look around for other goodies.) Make sure to set the ORAUSER, ORAPASS, and ORATNS constants to values appropriate to your system. (Oh, and you may not have a SCOTT.EMP table. *Smiles*)
#!/usr/bin/perl -w
use strict;
#-- Use modules
use DBD::Oracle;
use DBI;
#-- Define local constants
use constant TRUE => 1;
use constant FALSE => 0;
use constant ORAUSER => 'my_ora_user';
use constant ORAPASS => 'secret_password';
use constant ORATNS => 'ora_db_tns_name';
#-- Define local variables
my $gDBHandle;
my $gSQLCmd;
my $gSQLHandle;
my $gTabName;
my $gTabSchema;
my @gFields;
#-- Initialize local variables
$gSQLCmd = 'SELECT column_name, ' .
' nullable, ' .
' data_type, ' .
' data_length ' .
' FROM dba_tab_columns ' .
' WHERE owner = ? ' .
' AND table_name = ? ' .
' ORDER BY column_id ';
$gTabName = 'EMP';
$gTabSchema = 'SCOTT';
#-- Connect to the database
$gDBHandle = DBI->connect
(
'dbi:Oracle:' . ORATNS, ORAUSER, ORAPASS,
{
AutoCommit => FALSE,
PrintError => FALSE,
RaiseError => FALSE,
}
) || die 'Could not connect to Oracle ['.$DBI::errstr.' - '.$DBI::er
+r.']';
#-- Get the data
$gSQLHandle = $gDBHandle->prepare($gSQLCmd)
|| die 'Error with SQL statement ['.$DBI::errstr.' - '.$DBI::err.']'
+;
$gSQLHandle->execute($gTabSchema, $gTabName)
|| die 'Error with SQL statement ['.$DBI::errstr.' - '.$DBI::err.']'
+;
while (@gFields = $gSQLHandle->fetchrow_array)
{
print $gFields[0],"\t",$gFields[1],"\t",$gFields[2],"\t",$gFields[3]
+,"\n";
}
#-- Close the database connection
$gDBHandle->disconnect();
#-- Exit
exit;
#-- End of Example
| [reply]
[d/l] |
DBI::Shell has some excellent code in it that I use for one of my utility scrips sub db_describe available in the newer distributions of DBI
| [reply]
[d/l]
[select] |
This doesnt seem to work for me. I think select top only
works for SQL Server, and I am using Oracle. Any other suggestions? Thanks! Robert
| [reply] |
Just drop the TOP 1 from the SQL. I added that to prevent the DB from selecting all records and doing to much work. However, having the $sth->finish call at the end should minimize the overhead by merely finishing up the statement handle without fetching the subsequent rows.
Of course, Oracle should support some form of the TOP syntax, but I don't know Oracle.
Cheers,
Ovid
Join the Perlmonks Setiathome Group or just click on the the link and check out our stats.
| [reply]
[d/l] |
| [reply]
[d/l]
[select] |
You can try the NAME and the TYPE DBI statement handle objects.
Here's how you can grab column names:
my $names = $sth->{NAME};
#$names is an array reference
for (@$names) {
print $_, "\n";
}
Here's how you can grab the datatype of the columns:
my $types = $sth->{TYPE};
#$types is an array reference
for (@$types) {
print $_, "\n";
}
The last example returns an array of integer values, the value indicates the datatype (see ANSI X3.135 for a description).
| [reply]
[d/l]
[select] |
Two minor notes:
First, you might want to be careful about the values returned by $sth->{NAME} if you care about case-sensitivity. The DBI doesn't make them consistent or modify them at all -- they're whatever the database returns. And what the database returns can be vendor specific. The NAME_uc and NAME_lc properties do what you'd expect.
Second, I think the integer values returned from $sth->{TYPE} correspond to the DBI SQL type constants. So you can do:
use DBI qw( :sql_types );
...
my $types = $sth->{TYPE};
foreach my $type ( @{ $types } ) {
print "VARCHAR" if ( $type == SQL_VARCHAR );
print "DATETIME" if ( $type == SQL_DATETIME );
print "BOOLEAN" if ( $type == SQL_BIT );
print "FLOAT" if ( $type == SQL_FLOAT );
...
}
Conveniently, most of them also map to the ODBC CLI specifications. :-)
You can find the type-to-constant mapping in the 'dbi_sql.h' header file found in your architecture-dependent perl library tree -- for example, mine is in site_perl/5.6.1/i686-linux/auto/DBI/dbi_sql.h
Chris
M-x auto-bs-mode | [reply]
[d/l]
[select] |
desc is a sqlplus command
You don't need DBI to do what you want
#!/usr/bin/perl -w
use strict;
my $cmd="sqlplus scott/tiger\@whatever";
open SQLPLUS, "|$cmd" or die "Ah poo!:$!\n";
print SQLPLUS<<DOIT;
desc user_tables
DOIT
close SQLPLUS;
| [reply]
[d/l] |