Question

I made a database which collects information on a daily basis. as like calender the database stores members’ daily amount. If month doesn’t match any existing month(M-Y) the database will create a new html month table. I have solved this problem, as follows:

mysql_query("something goes here");
while(condition)
{
    mysql_query("something goes here")
    while(condition)
    {
        mysql_query("something goes here");
        while()
        {
            ....................
        }
        ........................
    }
}

This algorithm worked well when I discovered it. However, after a few days, it was placing a heavy load on my server. I then tried the same algorithm in PHP (but I can’t this). How can I make this run faster?
The code is as follows:

        $q2=mysql_query("SELECT  a.member_id,a.dates,MONTH(dates) AS months,
                        YEAR(dates)AS years,sum(amount) as sums
                        FROM account AS a
                        left join member as m
                        on(a.member_id=m.member_id)
                        GROUP BY (SELECT EXTRACT(YEAR_MONTH FROM dates))
                        ORDER by dates DESC
        ");
        $k=0;
        while($data2=mysql_fetch_assoc($q2))
        {
            $months=$data2['months'];
            $years=$data2['years'];
            $daten = new DateTime($data2['dates']);

            print "<tr><th align='left'><b>".$daten->format('F-Y')."</b></th>";

            $q3=mysql_query("select * from member");

            while($data3=mysql_fetch_assoc($q3))
            {
                $id3=$data3['member_id'];
                $q4=mysql_query("
                                SELECT SUM(amount) AS total FROM account
                                WHERE member_id=$id3
                                AND month(dates)=$months
                                AND year(dates)=$years                              
                            ");
                while($data4=mysql_fetch_assoc($q4))
                {
                    $total=$data4['total'];

                    print "<td class='total'>".number_format($total)."</td>";
                }
            }
            print "<td class='total'><b>".$data2['sums']."</b></td></tr>";
            $k=$k+$data2['sums'];
        }

Answer 1

Among other things:

• You’re running the query SELECT * FROM member for every row in the first query. This query is independent of the loop, so running it again every time is wasteful.

• For each result from the SELECT * FROM member query, you’re running another query (SELECT SUM(amount) AS total FROM account ...). There are several issues with this query:

  • First of all, this query could be combined into the previous query using a GROUP BY, to avoid having to run one query for every member. Something like:

    SELECT member_id, SUM(amount) AS total FROM account WHERE ... GROUP BY member_id

  • Second of all, you’re using MONTH(dates) = $months AND YEAR(dates) = $years. This is inefficient, as it forces the server to examine every row; converting it to a range on dates (e.g, dates BETWEEN '$year-$months-01' AND '$year-$months-31') would speed things up if there were an appropriate index on dates.

In general: Avoid queries in loops. The number of queries involved in generating a page should, to the degree possible, always be a small, nearly constant number. It should not grow with your data.