Question

I am having trouble with a function that checks if a set of user entered info (username and password) exists within either of the two possible tables where this information is stored.

The first table is the users table. It contains the first set of specific user information.

The last table is the listings table. It contains the second set of specific user information.

I have basically modified my original code to include the new listings table, and hence the trouble coming from within that task. The old code basically counted the number of results in the users table, if the result was greater than 0, then the function returned true, else false.

Now I have been stuck on the best way to go about adding another table to the query, and function. So I have been playing around with a union.

This was the original query:

SELECT COUNT(*) FROM users
WHERE id='$accNum' AND password='$password'

This returned a count of either 0 or 1 based on the info stored in the users table.

This is how I have reworked the query to include a count of the additional listings table:

SELECT count . *
    FROM (
        SELECT COUNT( * )
        FROM users
        WHERE id = '$accNum'
        AND PASSWORD = '$password'
        UNION (

        SELECT COUNT( * )
        FROM listings
        WHERE id = '$accNum'
        AND PASSWORD = '$password'
        )
    )count

This returned a result set of two rows, the first relating to the users table, and the second relating to the listings table. Then a column called COUNT (*) that contained the result count. This is the result set that I see within php myadmin.

Now this is the function:

function databaseContainsUser($accNum, $password)
{
    include $_SERVER['DOCUMENT_ROOT'] . '/../../includes/db.inc.php';

    $accNum = mysqli_real_escape_string($link, $accNum);
    $password = mysqli_real_escape_string($link, $password);


$sql = "SELECT count . *
        FROM (
            SELECT COUNT( * )
            FROM users
            WHERE id = '$accNum'
            AND PASSWORD = '$password'
            UNION (

            SELECT COUNT( * )
            FROM listings
            WHERE id = '$accNum'
            AND PASSWORD = '$password'
            )
        )count
";

    $result = mysqli_query($link, $sql);
    if (!$result)
    {
        $error = 'Error searching for user.';
        include 'error.html.php';
        exit();
    }

    $row = mysqli_fetch_array($result);

    if ($row[0] > 0)
    {
        return TRUE;
    }
    else
    {
        return FALSE;
    }
}

The problem that I have, is trying to work out how exactly to check the results to ascertain if the given log in credentials are valid.

I tried this: if (($row[0] > 0) || ($row[0] > 0)) But a var dump on $row showed that only the first row (count of users table) was being added to the array.

So I decided that this was complicated, and a long way to the final result.

So I tried selecting only the id column of the result as in:

...
`COUNT( * )` to `id`
...
$data = mysql_query($sql);
$num_sql = mysql_num_rows($data);
if ($num_sql > 0)
...

But this did not work out for me either.

But in either instance, my hours of trial and error have provided me with no success… So I’ve decided to seek help from the knowledgeable members of Stack Overflow!

So my question is this, what would be a logical way of going about this task? I am looking for any suggestions, or positive input what so ever here.

As I am fairly new to dabbling with PHP and mysql, if you would like to provide some code to explain your suggestions or input on the matter, it would more than likely help me to better understand the answer.

Answer 1

If you are checking existence only try doing this that way:

select case when 
    exists (SELECT 1 FROM users WHERE id = '$accNum' AND PASSWORD = '$password') or 
    exists (SELECT 1 FROM listings WHERE id = '$accNum' AND PASSWORD = '$password')
    then 1 else 0 
end as itDoesExist

It returns always one row with one column with 1 when record exists in at last one table (else 0).

Do not use count to check whether some specific record/-s exist/-s in table, it’s usually slower than simple exists.