note Revelation I'm going to go through some things that I thought were lacking in this tutorial. They aren't necessary, but are helpful. Selectrow_array: One thing many perl coders try to do is make their code as short as possible. One way to shorten the code is to use selectrow_array, instead of fetcrow_array. This is helpful if you are only executing a query once, and only getting one row of data. <CODE> my @row = $dbh->selectrow_array(q{SELECT my_data1, my_data2 WHERE furniture_type=?},undef, $furniture_type); </CODE> This is an extremely simple, and elegant way of selecting, and will be as optimized as any selectrow, that is executed without binding, as $furniture_type is the data executed. RaiseError: I find RaiseError to be extremely helpful for mission critical queries, and the majority of queries are mission critical to a script. Therefore it may be easyer to just set RaiseError, and stop using <code>or die</code>. <CODE> $dbh = DBI->connect($DBDSN, $DBUser, $DBPassword, {RaiseError => 1}); </CODE> To unset RaiseError on a query in which you want to specify something else to do, if the query doesn't work just add the code: <code>$dbh->{RaiseError} = 0;</code>, and set raiserror back to 1, after you are done. Bind_columns: <CODE> my ($furniture_name, $price); chomp($furniture_type); $sth->execute($furniture_type) or die "Couldn't execute: '$DBI::errstr'"; $sth->bind_col(1, \$furniture_name); $sth->bind_col(2, \$price); </CODE> You don't need to bind a specific column here. Instead you can just bind all the columns in order, with bind_columns. So instead of all that code you can write: <CODE> chomp($furniture_type); $sth->execute($furniture_type) or die "Couldn't execute: '$DBI::errstr'"; $sth->bind_columns(\my ($furniture_name, $price)); </CODE> This gets rid of predeclaring the variables, and using bind_col twice. Update: Did some research yesterday and found that bind_columns is faster for *all* queries! Strange, but it seems that bind_columns is faster than a fetchrow_arrayref, even if only one row is being caught. This renders fetchrow_array(ref) obselete in a sense, without using bind_columns, as a few lines of code makes a query much faster, regardless of how many rows are being returned. You may wish to check out <a href="http://archive.develooper.com/dbi-dev@perl.org/msg00049.html">this</a> to see the benchmarking, and query analysis that shows that. Finishing Queries: One thing that is rather interesting is the code: <Code>$sth->finish</Code>! Most tutorials don't teach users to finish their queries, as it's not really necessary, in the majoiry of situation. I would advise reading the finish section of perldoc DBI, as finish is pertinant in some situations. A link to that is <a href="http://www.perldoc.com/perl5.6.1/lib/DBI.html#finish">here</a>, as there's no reason for me to describe what has so aptly been described by Carlos Ramirez. It's really code optimization, but since you went into bind_col, I figured some more optimization would be helpful. Would like to commend the writer of this tutorial on the vast amount of information! Great job. <hr> Gyan Kapur gyan.kapur@rhhllp.com 7563 7563