Question

I’m currently trying to update my in house search engine to use multiple words. It’s very big and complicated but I’m running into errors I never ran into before now and I have no idea why.

so the question is, why am I receiving the error below?

I’m going to split this up into different sections for better understanding.

This is just the SQL query that I echod, copy and pasted into PHPMyAdmin, pasted here (PHPMyAdmin formatted it nicely) It searches for 2 words in the database:

SELECT *
FROM (

SELECT p.page_url AS url, COUNT( * ) AS occurrences
FROM PAGE p, word w, occurrence o
WHERE (
(
p.page_id = o.page_id
AND w.page_word_id = o.page_word_id
AND w.word_word LIKE '%' 'test' '%'
GROUP BY p.page_id
)
OR (
p.page_id = o.page_id
AND w.page_word_id = o.page_word_id
AND w.word_word LIKE '%' 'search' '%'
GROUP BY p.page_id
)
UNION (

SELECT f.file_url AS url, COUNT( * ) AS occurrences
FROM files f, filenames fn, fileoccurrence fo
WHERE f.file_id = fo.file_id
AND fn.file_word_id = fo.file_word_id
AND fn.file_word LIKE '%' 'test' '%'
GROUP BY f.file_id
)
OR (

SELECT f.file_url AS url, COUNT( * ) AS occurrences
FROM files f, filenames fn, fileoccurrence fo
WHERE f.file_id = fo.file_id
AND fn.file_word_id = fo.file_word_id
AND fn.file_word LIKE '%' 'search' '%'
GROUP BY f.file_id
)
)t
ORDER BY occurrences DESC 

This code is produced by this PHP code below which uses the explode function from the search input

// Do a little formatting
$keyword = strtolower($keyword);

// Get timestamp for start
$start_time = microtime(true);

$searched_words = explode(' ', $keyword);

foreach ($searched_words as $index => $word) {
    // Set up the stemmer
    $stemmer = new PorterStemmer;
    $stemmed_string = $stemmer->stem($word);
    $searched_words[$index] = $stemmed_string;
}

//  Configure the sql code
$sql = "SELECT * FROM (SELECT p.page_url AS url, COUNT(*) AS occurrences 
    FROM page p, word w, occurrence o WHERE (";

// Add the extra words to the sql
foreach ($searched_words as $index => $word) {  
    $sql .= "(p.page_id = o.page_id AND w.page_word_id = o.page_word_id
        AND w.word_word LIKE '%' '" . $word . "' '%' GROUP BY p.page_id) OR";
}
// Add the union to the sql and then add the second query   
$sql = substr($sql, 0, (strLen($sql)-3)); //this will eat the last OR
$sql .= " UNION ";

// The second set of querys
foreach ($searched_words as $index => $word) {
    $sql .= "(SELECT f.file_url AS url, COUNT(*) AS occurrences FROM files f, filenames fn, fileoccurrence fo
        WHERE f.file_id = fo.file_id AND fn.file_word_id = fo.file_word_id AND fn.file_word
        LIKE '%' '" . $word . "' '%' GROUP BY f.file_id) OR";
}

// Clsoe the sql code
$sql = substr($sql, 0, (strLen($sql)-3)); //this will eat the last OR
$sql .= ") t ORDER BY occurrences DESC"; // LIMIT " . $results . "");

// echo the query for the pure lolz of it
echo $sql . "<br /><br />";

// Search the DB for the results
$results = mysql_query($sql)
    or die("Invalid query: " . mysql_error());

All of this returns the error:

Invalid query: You have an error in your SQL syntax; check the manual
that corresponds to your MySQL server version for the right syntax to
use near ‘GROUP BY p.page_id) OR(p.page_id = o.page_id AND
w.page_word_id = o.page_word_id’ at line 3

Why am I receiving this error? The code I used previously worked fine. The only real thing I added was the foreach().

This is my original SQL code:

$result = mysql_query("SELECT * FROM (SELECT p.page_url AS url, COUNT(*) AS occurrences 
    FROM page p, word w, occurrence o WHERE p.page_id = o.page_id AND w.page_word_id = o.page_word_id
    AND w.word_word LIKE '%' '" . $stemmed_string . "' '%' GROUP BY p.page_id UNION  
    SELECT f.file_url AS url, COUNT(*) AS occurrences FROM files f, filenames fn, fileoccurrence fo
    WHERE f.file_id = fo.file_id AND fn.file_word_id = fo.file_word_id AND fn.file_word
    LIKE '%' '" . $stemmed_string . "' '%' GROUP BY f.file_id) t ORDER BY occurrences DESC") // LIMIT " . $results . "")
        or die("Invalid query: " . mysql_error());

---

EDIT: Fixed the above error. with this code

//  Configure the sql code
$sql = "SELECT * FROM (SELECT p.page_url AS url, COUNT(*) AS occurrences 
    FROM page p, word w, occurrence o WHERE (";

// Add the extra words to the sql
foreach ($searched_words as $index => $word) {  
    $sql .= "(p.page_id = o.page_id AND w.page_word_id = o.page_word_id
        AND w.word_word LIKE CONCAT('%', '" . $word . "', '%'))) OR "; //GROUP BY p.page_id)
}
// Add the union to the sql and then add the second query   
$sql = substr($sql, 0, (strLen($sql)-4)); //this will eat the last OR
$sql .= " GROUP BY p.page_id)";
$sql .= " UNION ";
$sql .= "(SELECT f.file_url AS url, COUNT(*) AS occurrences FROM files f, filenames fn, fileoccurrence fo
    WHERE (";

// The second set of querys
foreach ($searched_words as $index => $word) {
    $sql .= "(f.file_id = fo.file_id AND fn.file_word_id = fo.file_word_id AND fn.file_word
        LIKE CONCAT('%', '" . $word . "', '%'))) OR "; //GROUP BY f.file_id)
}

// Clsoe the sql code
$sql = substr($sql, 0, (strLen($sql)-4)); //this will eat the last OR
$sql .= " GROUP BY f.file_id)";
$sql .= ") t ORDER BY occurrences DESC"; // LIMIT " . $results . "");

This now produces error:

Invalid query: Every derived table must have its own alias

which has something to do with the unions (i’m not very good at unions.. or SQL in general)

Answer 1

You can’t concatenate strings in SQL just by placing them next to each other.

AND w.word_word LIKE '%' 'test' '%'

Should be

AND w.word_word LIKE CONCAT('%', 'test', '%')

Or if you use SET SQL_MODE='PIPES_AS_CONCAT' to get standard ANSI SQL syntax, you can use:

AND w.word_word LIKE '%' || 'test' || '%'

---

Re your comment, I see another problem:

In SQL, the WHERE clause must be complete before you can add a GROUP BY clause. The syntax is:

WHERE ( <conditions...> )
GROUP BY <expressions>

Whereas you have:

WHERE ( <conditions...> GROUP BY <expressions> ) 
   OR ( <conditions...> GROUP BY <expressions> )

What you have is not legal syntax.

Really, this is the kind of thing you should be able to look up yourself in any beginner reference on SQL.

---

Every derived table must have its own alias

This means that you used a subquery in the FROM clause but did not give it an alias. For example:

SELECT ... FROM (SELECT ... FROM table) AS x WHERE ...etc... 

If you leave out the AS x then it’s an error (x is just an example in this case, you can choose a more meaningful alias).