Randomization by Integer

[AlphaShade]

Hello!

Well basically I have a problem.

I have a table with a value: Popularity. What I want to do is do a site randomization.

I have rows: id, url, website_description, popularity

What I want to do is do a randomization where a site will come up in the script by a popularity number of 1 to 100(or more).

My problem is I have seen plenty of randomization scripts out there but nothing this complex.

Wondering if someone knows a way to do this. I want it to randomize all sites but those sites that have the highest popularity rating would show up more then the sites with the lowest.

Any help you can give me would be great!
Thanks for reading 🙂

[Installer]

This function, posted in the user notes for array_rand(), works pretty well (unchanged except for formatting):

function array_rand_weighted($values) 
{
    $r = mt_rand(1, array_sum($values));
    foreach ($values as $item => $weight) {
        if  ($r <= $weight) {
            return $item;
        }
        $r -= $weight;
    }
}

You could use it something like this:

$query = "SELECT url, popularity FROM my_table";
$result = mysql_query($query);
while ($row = mysql_fetch_array($result)) {
    $array[$row['url']] = $row['popularity'];
}

$num_links = 10;
$ct = 0;
$links = array();
while ($ct < $num_links) {
    $tmp = array_rand_weighted($array);
    if (in_array($tmp, $links)) {
        continue;
    } else {
        $links[$ct++] = $tmp;
    }
}

[AlphaShade]

Thanks for the reply!

This is what I have:

// Connect to the database
mysql_connect ('localhost', 'site_testdb', 'XXXXXX') ;
mysql_select_db ('site_testdbuser')or die(mysql_error()); 

echo "1";
$query = "SELECT url, popularity FROM linkstable";
$result = mysql_query($query);
while ($row = @mysql_fetch_array($result)) {
    $array[$row['url']] = $row['popularity'];
}

function array_rand_weighted($values) {
   $r = mt_rand(1, array_sum($values));
   foreach ($values as $item => $weight) {
       if  ($r <= $weight) return $item;
       $r -= $weight;
   }
}


$num_links = 10;
$ct = 0;
$links = array();
while ($ct < $num_links) {
    $tmp = array_rand_weighted($array);
    if (in_array($tmp, $links)) {
        continue;
    } else {
        $links[$ct++] = $tmp;
    }
}

I had to put the @ in front of mysql_fetch.

I get the following error:
Fatal error: Call to undefined function: array_rand_weighted() in /query.php on line 19

So Basically that means I don't have that above function declared and I put it in.

So I then end up with this:

// Connect to the database
mysql_connect ('localhost', 'site_testdb', 'XXXXXX') ;
mysql_select_db ('site_testdbuser')or die(mysql_error()); 

echo "1";
$query = "SELECT url, popularity FROM linkstable";
$result = mysql_query($query);
while ($row = @mysql_fetch_array($result)) {
    $array[$row['url']] = $row['popularity'];
}

function array_rand_weighted($values) {
   $r = mt_rand(1, array_sum($values));
   foreach ($values as $item => $weight) {
       if  ($r <= $weight) return $item;
       $r -= $weight;
   }
}


$num_links = 10;
$ct = 0;
$links = array();
while ($ct < $num_links) {
    $tmp = array_rand_weighted($array);
    if (in_array($tmp, $links)) {
        continue;
    } else {
        $links[$ct++] = $tmp;
    }
}

And then I get this:

Warning: array_sum(): The argument should be an array in /query.php on line 16

Warning: Invalid argument supplied for foreach() in /query.php on line 17

So you may be smacking your face because I am a noob. Any Idea why it is doing that? I tried to edit the code around but still come up with the error.

My Goal is to have 1 link be spit out as well not a bunch of links.

The goal is to have a redirection to the url in the table so I would take a redirection snippit and input the $randomization_result into that and then have it open up a new page with that.

Thank you for your time 🙂 Sorry for being such a newbie.

[Installer]

...I don't have that above function declared and I put it in.

Of course. User functions tend not to work unless they're put in. That's why I reformatted and posted it for you earlier.

I had to put the @ in front of mysql_fetch.

Why? I suspect that's the answer to the problem of:

The argument should be an array...

Your query isn't returning any results. Check your database.

...1 link...not a bunch of links.

Then don't loop. My example code was only...an example. So change this:

$num_links = 10;
$ct = 0;
$links = array();
while ($ct < $num_links) {
    $tmp = array_rand_weighted($array);
    if (in_array($tmp, $links)) {
        continue;
    } else {
        $links[$ct++] = $tmp;
    }
}

to this:

$link = array_rand_weighted($array);

You need to find out why "$array" isn't being filled. Here are a couple of hints:
- mysql_num_rows()
- mysql_query() returns false on failure.

Good luck and keep us posted.

[AlphaShade]

Ok I found out the problem. My naming of the table rows. What was referenced in the example didn't exactly reflect what was in my table and I forgot to rename.

So now it spits out one link and seems to be working.

So I come to my next problem. The num links portion seems to must be exactly the number of rows in that table. Is it possible to have num links dynamically populate the number of current rows without having to go into the code each time and put in how many are actually there?

For instance I have 3 now and I must put 3. If i have 10 I will have to go back into it and put 10.

Is there a way to return with the current number of rows for that table?

Thanks for the help so far.

[Installer]

In my code, "$num_links" is merely the number of times the loop will iterate; i.e. it's the number of random urls that will be returned in an array. It has nothing to do with the number of rows returned by the query (except it has to be smaller). The code will work with any size result > 0 and >= $num_links.

I'm guessing that's what you mean, but, frankly, I'm not sure it is. Anyway, hth.

[AlphaShade]

Ah I see. It was set to 10 and the page would basically never load. I set it to 3 (the current number of rows) and it loaded fine.

So num_links itself has nothing to do with the number or rows it looks through. It will still randomize the entire set of rows?

Pretty interesting how it must be at or less then the number or rows for it to work. Wondering why that is.

And by rows I mean current data entered into the table. Rows of data. There is 3 entries right now.

[Installer]

So num_links itself has nothing to do with the number or rows it looks through. It will still randomize the entire set of rows?

The function looks through the entire array each time it's called. "$num_links" is merely the number of times the function is called. It can be set statically, dynamically, any way and to whatever value you want. If you set it too high it won't work right, like you found out. But setting it to the same size of the array would be pointless; it'll just result in every element being returned.

...it must be at or less then the number or rows for it to work. Wondering why that is.

? It would be the same as having two scoops of ice-cream and asking for a triple-decker.

I've got an idea for speeding the process up, especially for a larger set of data. I'll post the code later.

[AlphaShade]

I see.

And I guess I am not entirely understanding thus the reason You are wondering what I mean.

If I were to set the num_links to 4 since there are 3 rows of data, the page will never load. If I add in a 4th row of data, then the page will load and that randomization will complete.

So To me it looks like to me that what is happening is that num_links is the number of rows it is displaying and then randomizing and then sorting by weight.

Of course as you said this is not true it is just the number of times the randomization is occuring...

I am wondering why exactly if I put it 1 number above how many rows of data there is currently that the page just never loads. Its not really a matter of the fact that it is randomizing alot, If I Put in 100 rows of data right now It probably would load fine if I put num_links as 100.

Ill try that to make sure.

I tried deleting all rows just adding one and putting num_links as 2 and it never loads.

I probably am missing your point entirely and that is why i am confused.

[Installer]

The reason the page never loads with $num_links greater than number of rows is that there is no validity checking on $num_links in the code, or any oher error-checking. Putting the code into production without that would not be a good idea. It's just bare-bones right now.

[AlphaShade]

Well What I Whipped up is a variable in place of that that counts the rows and puts it as how many current rows there are.

// Connect to the database
$link = mysql_connect ('localhost', 'site_user', 'XXXX') ;
mysql_select_db("site_database", $link)or die(mysql_error()); 

$result = mysql_query("SELECT * FROM table", $link);
$num_rows = mysql_num_rows($result);

$query = "SELECT url, popularity FROM table";
$result = mysql_query($query);
while ($row = mysql_fetch_array($result)) {
    $array[$row['url']] = $row['popularity'];
}


function array_rand_weighted($values) {
   $r = mt_rand(1, array_sum($values));
   foreach ($values as $item => $weight) {
       if  ($r <= $weight) return $item;
       $r -= $weight;
   }
}


$num_links = $num_rows;
$ct = 0;
$links = array();
while ($ct < $num_links) {
    $tmp = array_rand_weighted($array);
    if (in_array($tmp, $links)) {
        continue;
    } else {
        $links[$ct++] = $tmp;
    }
}

echo "$tmp <br>";

echo "$num_rows Rows\n";

And that seems to work dynamically. However you said

But setting it to the same size of the array would be pointless; it'll just result in every element being returned.

I am wondering what you mean by this. I do infact want everything being returned from the table so that it will randomize the entire table and pick out the link that appears at the top which I am assuming it is currently doing as I return 1 link instead of spitting out the entire table.

I think I understand what you mean because Your code looks to supposed to spit out 10 Links sorted by popularity. Right now it spits out only 1. Thats why I think you said then don't loops so then it doesnt spit out more then one.

As You can see I am echoing $tmp as it seems to be the only one that shoots out the result. Therefore that coule be the reason why only 1 is being returned because I am echoing the wrong result string.

Now that I look at it I think thats what you meant.

I am sorry if you are getting frustrated in trying to explain I really do appreciate your help!

[Installer]

Well, I guess I misunderstood when you said this:

My Goal is to have 1 link be spit out as well not a bunch of links.

"$tmp" is merely that--a temporary variable. The urls are in the "$links" array. Sorry, but I thought that was evident. To display the urls, then, do:

foreach ($links as $url) {
    echo $url . '<br />';
}

instead of "echo $tmp;".

Edit: My idea for new code won't work.

You're right, returning all the links should work well.

[AlphaShade]

Its cool I blame myself not being clear enough because I am a bit new to sql and php.

That appears to work though everything seems to be working great. Thank you very much for your help!