things just arnen't in order!

tgmsocal

I have a search engine script with ranking. The script counts the amount of times the search terms appear in the description, and then sorts the results by the amount of matches.

The problem is, things aren't coming up in order everytime. The results are grouped by their ranking and print out correctly, but the order isn't ever right.

If you need to see more than the code below, just ask!

Thanks for any help,

tgmsocal


// count word matches and rank the listing accordingly
while($row = mysql_fetch_array($result, MYSQL_ASSOC)) {
	$row[score] = count_words($row[video_desc]);
	$rows[] = $row;
}

uasort($rows, 'search_rank');

// define which rows to print out of ranked results
$rowtoprintstart = ($s);
if(($s + 9) <= ($numrows)){
	$rowtoprintend = ($s + 9);
}
else {
	$rowtoprintend = (($numrows - $s) + ($rowtoprintstart - 1));
}

// then, print the results!
for($rowtoprint = $rowtoprintstart; $rowtoprint <= $rowtoprintend; $rowtoprint++) {
	$rowarray = $rows[$rowtoprint];
	$title = $rowarray["video_name"];
	$desc = $rowarray["video_desc"];
	$path = "/h1/adamspro/".$rowarray["video_path"];
	$category = $rowarray["video_cat"];
	$id = $rowarray["id"];
	if($rowarray["score"] == 1) {
		$score = "<font face=\"Arial, Verdana\" size=\"1\">".$rowarray["score"]." WORD MATCH</font>";
	}
	else {
		$score = "<font face=\"Arial, Verdana\" size=\"1\">".$rowarray["score"]." WORD MATCHES</font>";
	}
	$fsize = filesize_mb($path);
	$prettydesc = highlight_words($desc);

echo("<br><br><b>$count.&nbsp;<a href=\"http://www.XXXXXXXXXX.com/videodat/video.php?id=$id\" target=\"_blank\">$title</a></b>&nbsp;($fsize)&nbsp;($score)&nbsp;<br>$prettydesc&nbsp;");
$count++;
}

// ranking function for use in uasort
function search_rank($a, $b) {
	$ax = $a[score];
	$bx = $b[score];
	if($ax == $bx) {
		return 0;
	}
	return ($ax > $bx) ? -1 : 1;
}

// count word matches function

function count_words($content) {
	global $searchwords;
	$wordcount = 0;
	$lowercasecontent = strtolower($content);
	foreach($searchwords as $word) {
		$firstword = explode(" ", $lowercasecontent, 2);
		if((strcasecmp($firstword[0], $word) == 0)) {
			$wordcount++;
		}
		$wordcount += substr_count($lowercasecontent, " ".$word." ");
		$wordcount += substr_count($lowercasecontent, " ".$word.".");
		$wordcount += substr_count($lowercasecontent, " ".$word."!");
		$wordcount += substr_count($lowercasecontent, " ".$word."?");
		$wordcount += substr_count($lowercasecontent, " ".$word.",");
	}
	return $wordcount;
}

Weedpacket

Well, considering that we don't have your database, or its schema (so that we can't, e.g., suggest whether using SUM(), GROUP BY and ORDER BY to gather the rankings in order originally) we certainly aren't in a situation to test anything. So right now it could be pretty much anything in that huge wodge of code that's causing the problem. If you could try and narrow things down and show something that (a) shows the problem, and (b) doesn't distract with irrelevancies, then that might help focus matters.

PS: Eeww. <font> tags! Thppt! thppt!

tgmsocal

Here is the database accessing code. I use RLIKE statements and I don't do an ORDER BY because the uasort command should sort the rows, which is why I only showed that command.

All the code that goes along with making the ranking score, sorting it, and displaying it is in that code, which is where the error must be. The count_words function generates a score, then the score is put into the row's array, then the array is sorted by the "search_rank" function.

I have out put the results of uasort and the results appear to be in the correct order. I can't figure out why the results continue to print up in incorrect order..

Thanks again, hope this script below helps you figure out where I went wrong!

--tgmsocal

(What's wrong with font tags?? 🙂 )

//connect to your database
if(!mysql_connect($server, $dat_user, $dat_pwd)) {
	die("Unable to connect to MySQL server" . mysql_error());
}

//specify database
@mysql_select_db($database) or die("Unable to select database" . mysql_error());

// explode search term into array, count words
$exp_var = explode(" ", $var);
$wordcount = count($exp_var);

if($wordcount > 1) {
	$query = "SELECT * FROM ($dat_table) WHERE (video_desc) RLIKE ('[[:<:]]".$exp_var[0]."[[:>:]]')";
	$crtwrd=2;

while($crtwrd <= $wordcount) {	
	$i = ($crtwrd - 1);
	$query .= " OR (video_desc) RLIKE ('[[:<:]]".$exp_var[$i]."[[:>:]]') OR (video_name) RLIKE ('[[:<:]]".$exp_var[$i]."[[:>:]]')";
	$crtwrd++;
}
}
elseif($wordcount == 1) {
	$query = "SELECT * FROM ($dat_table) WHERE (video_desc) RLIKE ('[[:<:]]".$trimmed."[[:>:]]')";
}
else {
	die("Script error! Unable to create database query! Contact webmaster for support.");
}

tgmsocal

I think there is something wrong with my uasort..

What I do is count the word matches and add them up, then stow them in the row as 'score'..

Then I do a uasort on the rows with the sort function looking at the 'score' of each row.

However, when I print out my results, the order is never right. It seems almost random on the first page, then least to most result matches on the 2nd page.. Obviously I'm looking to list most to least over multiple pages..

here is the specific code in question:


while($row = mysql_fetch_array($result, MYSQL_ASSOC)) {
	$row[score] = count_words($row[video_desc]);
	$rows[] = $row;
}

uasort($rows, "search_rank");

function search_rank($a, $b) {
	$ax = $a[score];
	$bx = $b[score];
	if($ax == $bx) {
		return 0;
	}
	return ($ax > $bx) ? -1 : 1;
}

There is the scoring process and sorting.. below is how it is all printed out:

for($rowtoprint = $rowtoprintstart; $rowtoprint <= $rowtoprintend; $rowtoprint++) {
	$rowarray = $rows[$rowtoprint];
	$title = $rowarray["video_name"];
	$desc = $rowarray["video_desc"];
	$path = "/h1/adamspro/".$rowarray["video_path"];
	$category = $rowarray["video_cat"];
	$id = $rowarray["id"];
	if($rowarray["score"] == 1) {
		$score = "<font face=\"Arial, Verdana\" size=\"1\">".$rowarray["score"]." WORD MATCH</font>";
	}
	else {
		$score = "<font face=\"Arial, Verdana\" size=\"1\">".$rowarray["score"]." WORD MATCHES</font>";
	}
	$fsize = filesize_short($path);
	$prettydesc = highlight_words($desc);

echo("<br><br><b>$count.&nbsp;<a href=\"http://www.XXXXXXXXX.com/videodat/video.php?id=$id\" target=\"_blank\">$title</a></b>&nbsp;($fsize)&nbsp;($score)&nbsp;<br>$prettydesc&nbsp;");
$count++;
}

I don't think the display code has anything to do with the problem, but if I knew what the problem was I wouldn't be asking this question on the forums 🙂.

Thanks for any help anyone can provide.. This is the first real wall I've run into so far with PHP!

--tgmsocal

Weedpacket

Just mentioning this in the meantime (too busy right now to pay much attention):

Originally posted by tgmsocal
(What's wrong with font tags?? 🙂 )

Apart from being deprecated since last century?

HTML 4.0 Specification; W3C Working Draft 8-July-1997
Font modifier elements: FONT and BASEFONT

FONT and BASEFONT are deprecated.

HTML 4.01 Specification; W3C Recommendation 24 December 1999
15.2.2 Font modifier elements: FONT and BASEFONT
FONT and BASEFONT are deprecated.