You will have to provide more information if you want a relevant response here at perlmonks. The speed of your database inserts generally depends on:

• Your hardware configuration and environment.
• Your disk system(s).
• Your database engine's configuration.
• The structure of your database instance (definition of tables, indices, and objects).
• The nature of the data you are inserting.
• The client script you use to access your database.

Without knowing any of the above there is no Monk in the world that can answer you.

Thank you. I add more information:

    * Your hardware configuration and environment.
    	My client host :192.168.11.12 ,and the SqlServer DB server:192.168.11.16 .They located at some place. 
    
    * Your disk system(s).
      My client host HD:300G. 
    
    * Your database engine's configuration.
    	DB is charged by our DBA.I really don't know what's config you means
    	
    * The structure of your database instance (definition of tables, indices, and objects).
    	Table Definition:
    	test_c (ID int not null,cust_id int not null,status char(1) null,money decimal(12,2) null)
      no indices 	
    
    * The nature of the data you are inserting.
      My data is from my client host's sqlite DB  
  
    
    * The client script you use to access your database. 
    

My Code :


     #!/usr/bin/perl
     use strict;
     use warnings;
     use DBI;
    
    
     my $mydbh = DBI->connect( "dbi:SQLite:/root/loc.db","","",{AutoCo
+mmit => 0} ) or die "Error $DBI::errstr \n"; 
     my $dbh   = DBI->connect( "dbi:Sybase:server=192.168.11.16;databa
+se=customer",
                        "xxx",
                        "xxx",
                        {AutoCommit =>1})
                        or die $DBI::errstr;
      
    insert_it();
    $mydbh->disconnect;
    $dbh->disconnect;
    
    exit;
    
    ##########################
    sub insert_it() {
      my $r = get_data();
      for (keys %$f) {
        my $mo         = $r->{$_}->{money};
        my $status= $r->{$_}->{status}; 
            $dbh->do(qq{insert into test_c(custid,money,status) values
+ ($_,'$mo','$status')}) or die $DBI::errstr;
      }
    }
    
    sub get_data {
      my %data;
      my $sth = $mydbh->prepare(qq{select custid,money,status from tem
+p_test order by custid});
      my ($cid,$mon,$sta);
      $sth->execute() or die $DBI::errstr;
      $sth->bind_columns(undef, \$cid, \$mon, \$sta);
      while ( $sth->fetch() ) {
        $data{$cid} = {money => $bal, status => $sta};
      }
      return \%data;  
    }
[download]

I have tried the placeholder method,but the sqlserver 2K doesn't support it.

DB is charged by our DBA.I really don't know what's config you means

Then talk to her/him and get involved! You cannot ignore how the database is set up if you want high performance.

I have tried the placeholder method,but the sqlserver 2K doesn't support it.

I don't think that is true. Read Placeholders(?) in the prepare statements for executing DBD::Sybase queries.
--
Andreas