Don't you think there's a contradiction here?
I beleive the idea of altering the table structure on the fly is great, so you dont have to mess with mysql all the time, altering tables, or dropping them and create new ones...
(Bold is mine) You either think that altering the table is a great idea, or that it's a complete mess. And you're right in the second case.
What Fletch meant<stroke>, I think,</stroke> is that you should have a table with two columns: one for the name of the page, and the other for the associated counter. Something like this: +----------+---------+
| pagename | counter |
+----------+---------+
| home | 123 |
| page_1 | 22 |
| page_2 | 16 |
| credits | 2 |
| policy | 1 |
+----------+---------+
This way, each new page is just a new record in this table, which you can do with a simple INSERT.
Hey! Up to Dec 16, 2007 I was named frodo72, take note of the change! Flavio
perl -ple'$_=reverse' <<<ti.xittelop@oivalf
Io ho capito... ma tu che hai detto?
| [reply] [d/l] [select] |
Yes indeed this is very clear and better.
But in order to increase a counter for a specific page or create it first and then populate it, the script must know which webpage just loaded for example page_5.html and extract just the filename without the dot and extension.
How to do this? The best i can do is this:
$pagename = "something that tells me what si the name of the currnt sc
+ript running, just its filename";
if ( $pagename is an already actual field inside table 'counter' then
+)
{
$select = $db->prepare( "SELECT pagename, pagecounter FROM counter
+s" );
$select->execute;
}
else
{
$select = $db->prepare( "INSERT INTO counters (pagename, pagecount
+er) VALUES (?, ?)" );
$select->execute( $pagename, 1 );
my $pagecounter;
while( $row = $select->fetchrow_hashref )
{
$pagecounter += $row->{pagecounter} unless ( $row->{host} eq "\
+5;ίκος" );
}
| [reply] [d/l] |
Nothing prevents you from using the full file name in the "pagename" column :)
Regarding how to get the page name, you provided little insight about how a page name is "generated"; in any case, whatever your approach (the one suggested by Fletch, or the ALTER TABLE one), you have to find some way to decide which counter you want to increase. Which means that this is not a problem tied to this specific solution. If the page name is tied to the current script, anyway, you might benefit from the contents of either $0 (perlvar) or of __FILE__ (perldata), whatever applies best to your case.
I'd make the "pagename" column a primary key for the table, and just do something like this:
# Unconditionally try to create a new record.
eval {
$db->do('INSERT INTO counters (pagename, pagecounter) VALUES ($, 0)
+',
undef, $pagename);
};
# Unconditionally increase the counter
$db->do('UPDATE counters SET pagecounter = pagecounter + 1 WHERE pagen
+ame = ?',
undef, $pagename);
If the record for the given page does not exist, it's created and the counter is initialised to 0. If it already exists, the INSERT fails due to the fact that the pagename column is a primary key, and you can't have two records with the same primary key. In either case, after the eval you're sufficiently confident* that a record for the given page actually exists in the table.
After this, you increase the counter. If the record was just created, it is correctly bumped to 1. In the other case, it is simply increased, which is what you want.
No checks (the RDBMS does these for you), no race conditions... maybe a little dirt? Acceptable, IMHO.
* "sufficiently confident" means that the INSERT could fail for other reasons apart the field being there. If all the rest is working properly, than you can be confident that after the eval you'll have something to increase in that table.
Hey! Up to Dec 16, 2007 I was named frodo72, take note of the change! Flavio
perl -ple'$_=reverse' <<<ti.xittelop@oivalf
Io ho capito... ma tu che hai detto?
| [reply] [d/l] [select] |