Question

I want to have a php page that display multiple keywords (checkboxes) for example areas such as Narberth, Pembrokeshire, Carmarthenshire etc as well as other things.

So far i have created a way to do this and display the data separately in tables. What I’m having trouble with is showing the rows with multiple keywords associated with them first.

So if the user searched “Narberth” and “Pembrokeshire” the results would show any documents with both those keywords in first then anything with just “narberth” and just “pembrokshire” afterwords.

I listed my code for what I’ve worked out so far :

<form method="post">
<table border="1px solid" width="60%">
<tr>
<td><input type="checkbox" name="keywords[]" value="1">Narberth </td>
<td><input type="checkbox" name="keywords[]" value="2">James Brothers </td>
<td><input type="checkbox" name="keywords[]" value="3">Mabinogion </td>
</tr>
<tr>
<td><input type="checkbox" name="keywords[]" value="4">Pembrokeshire </td>
<td><input type="checkbox" name="keywords[]" value="5">Pubs </td>
<td><input type="checkbox" name="keywords[]" value="6">Coastal Paths </td>
</tr>
</table>
<br>
<input type="submit" name = "submit">

<?php
if (isset($_POST['submit'])) {

foreach( $_REQUEST['keywords'] as $keywordID )
{
$result = mysql_query("SELECT * FROM tbl_index m
inner join tbl_filekeywords r on m.file_id = r.file_id
where r.keyword_id = '".$keywordID."'");

$keywordname = mysql_query ("SELECT keyword FROM tbl_keywords WHERE keyword_id = '".$keywordID."'");
$KeywordName = mysql_fetch_row($keywordname);
$KeywordName = $KeywordName[0];

echo "<table border='1' width='50%'>
<th colspan='2'>".$KeywordName."</th>
<tr>
<th>File ID</th>
<th>Filename</th>
</tr>";

while($row = mysql_fetch_array($result)){
echo "<tr>";
echo "<td>" . $row['file_id'] . "</td>";
echo "<td>" . $row['file'] . "</td>";
echo "</tr>";
}
echo "</table><br>";
mysql_free_result($result);
}
}?>

P.S There is another alternative that i developed that lumps all the files into one table which is fine but i still want to stop the same files appearing multiple times for each keyword and would rather the ones with more than one keyword to be prioritized and displayed on top.

<?php
if(isset($_POST['keywords']) && !empty($_POST['keywords'])){
foreach($_POST['keywords'] as $key=>$value){
if($value==1) $keywords[] = "keyword_id='".mysql_escape_string($key)."'";
}
$keywords = implode(' OR ', $keywords);
}
$query = "SELECT * FROM tbl_index m inner join tbl_filekeywords r on m.file_id = r.file_id WHERE $keywords";
$result = mysql_query($query);

echo "<table border='1' width='50%'>
<tr>
<th>File ID</th>
<th>Filename</th>
<th>Keyword_ID</th>
</tr>";

while($row = mysql_fetch_array($result)){
echo "<tr>";
echo "<td>" . $row['file_id'] . "</td>";
echo "<td>" . $row['file'] . "</td>";
echo "<td>" . $row['keyword_id'] . "</td>";
echo "</tr>";
}
echo "</table><br>";
mysql_free_result($result);
?>

And finally the tables are set up like:

tbl_index : file_id, file

tbl_keywords : keyword_id, keyword

tbl_filekeywords : file_id, keyword_id

Answer 1

First of all, wellcome to SO. Post answer delay is due to non synthesized question. In next questions remember to include only relevant code to question.

Lets supose that tbl_index and tbl_filekeywords tables are in 1:N relationship.

First of all, in php you can get $keywords as a string list of selected keywords Example:

$keywords = " 1, 5, 12,  24"

Now you can rank tbl_filekeywords in this way. We named rank view:

select 
  r.file_id, 
  sum( case when r.keyword_id in ( $keywords  ) then 1 else 0 end ) as rank
from
  tbl_filekeywords r
group by 
  r.file_id

Last step is to join rank view with tbl_index:

   select t.*
   from 
     tbl_index t left outer join
     ( HERE PREVIUS QUERY ) r
          on r.file_id = t.file_id
   order by
     coalesce( r.rank, 0 )

Let us to know if this runs and it is a solution for you. If not, explain how can we improve answer!

Edited due OP Comment

I don’t have your database schema, but it seems that file_id is PK for tbl_index. If this is the case, the query don’t produce duplicates. Notice that first query, rank view, only has a row for each file_id due to group by clause. Because this, join don’t produce duplicates. For me the final query is:

   select t.*
   from 
     tbl_index t inner join                 <--No dups
     ( 
       select 
         r.file_id, 
         count(*) as rank
       from
         tbl_filekeywords r
       where
         r.keyword_id in ( $keywords  )
       group by 
         r.file_id                          <--No dups
     ) r2
          on r2.file_id = t.file_id
   order by
      r2.rank