$db->SetFetchMode() is other way to set this setting.

From Martin Dougiamas (martin at moodle.com) Friday, 15 October 2004, 07:51 PM:

It doesn't actually duplicate anything ... try this:

foreach ($USER as $key => $data) { echo <p>$key ... $data</p>; }

And it's not lengthening any foreach loops, because there's no reason to loop through the fields of a record ...

Best to leave it as is, I think.

From Petri Asikainen (paca at sci.fi) Sunday, 17 October 2004, 10:37 PM:

I find this out when I was writing user syncronisation and using code like:

$users = get_records(user, 'auth', 'ldap');

foreach($users as $user) {

$update='';

$count=0;

$numddd = count($user);

foreach ($user as $key => $value) {

$count++;

$update .= $key .' = \''. $value .'\''.\n;

if ($count < $numddd) { $update .= ', '; }

}

echo $update.\n\n\n;

}

Every update_record and insert_record (by using getInsertSQL) have to use loops handle data. But I havent tested is there any performance differences or not.

This function always assumes that the first field can be used as a key, so it uses it as a key and slices it off (array_slice($this->fields, 1) from the "value" part of the returned array. Moodle on the other hand very often believes that there is a field named "id" that has the id value in it.

This combination of behaviours works with *_BOTH because now ADODB uses the first field (now the numerical reference to the id field) as a key and then slices it off, still leaving the associative reference to the id field. This is why the ID field is the only non-duplicated field. I used to think that it's pre-pasted in there by Moodle itself, but it's a bit reverse: the dupe of it is removed by ADODB."

More at http://moodle.org/mod/forum/discuss.php?d=27897 and http://moodle.org/mod/forum/discuss.php?d=36219.