MySql Tally system with php - php

Question

I'm trying to condense data that I have in my database into rows with their points tallied to see the most popular.
If I had a data table like:
`data`
item1 item2
1
1 2
1 3
1 3
2 3
And wanted the condensed version to be:
`data_sum`
item1 item2 Tally
1 2 2
1 3 3
2 3 1
How would I achieve this? I have somewhat of an idea here:
$popdata = mysql_query("SELECT * FROM data");
while($add = #mysql_fetch_array($popdata)){
$qitem1 = "SELECT * FROM data_sum WHERE item1='".$add['item1']."'";
$ritem1 = mysql_query($qitem1);
if(mysql_num_rows($ritem1) > 0){
$qitem2 = "SELECT * FROM data_sum WHERE item2='".$add['item2']."'";
$ritem2 = mysql_query($qitem2);
if (mysql_num_rows($ritem2) > 0){
$sql = "UPDATE Tally=Tally + 1 WHERE item1='".$add['item1']."' AND item2='".$add['item2']."'";
$update = mysql_query($sql);
}
else{
$sql = "INSERT INTO data_sum (item1, item2) VALUES('$item1', '$item2')";
$insert = mysql_query($sql);
}
else{
$sql = "INSERT INTO data_sum (item1, item2) VALUES('$item1', '$item2')";
$insert = mysql_query($sql);
}
Yes, I know the total tallies are one more than the rows in the first table. I want the rows with a null column to count towards both tallies with a common factor. This file is going to go through thousands of rows so I want utmost efficiency! Thanks!

Answer 1

All you would need to do is create a new table and then combine an INSERT statement with a GROUP BY'd SELECT statement. This would COUNT() the number of times item1 and item2 were the same and store them in the new tally'd table.
Something along the lines of:
INSERT INTO new_tally_table (item1, item2, Tally)
SELECT item1, item2, COUNT(*)
FROM table
GROUP BY item1, item2
Edit:
Actually re-read the last bit of your question. Think what you want is something like this:
SELECT item1, item2, COUNT(*)
FROM (
SELECT i1.item1, i2.item2
FROM table1 as i1
INNER JOIN (
SELECT DISTINCT item1, item2
FROM table1 WHERE item2 IS NOT NULL
) as i2 ON (i1.item1 = i2.item1)
WHERE i1.item2 IS NULL
UNION ALL
SELECT item1, item2
FROM table1
WHERE item2 IS NOT NULL
) as t
GROUP BY item1, item2
There's probably a better way of writing that though.

Answer 2

There may be a simpler solution, but I can't think of it...
DROP TABLE IF EXISTS my_table;
CREATE TABLE my_table
(id INT NOT NULL AUTO_INCREMENT PRIMARY KEY
,item1 INT NULL
,item2 INT NULL
);
INSERT INTO my_table (item1,item2) VALUES
(1 ,NULL),
(1 ,2),
(1 ,3),
(1 ,3),
(2 ,3);
SELECT x.item1
, x.item2
, COUNT(DISTINCT y.id)
FROM my_table x
JOIN my_table y
ON y.item1 = x.item1
AND (y.item2 = x.item2 OR y.item2 IS NULL)
AND y.id <= x.id
JOIN
( SELECT item1
, item2
, MAX(id) max_id
FROM my_table
GROUP
BY item1
, item2
) z
ON z.item1 = x.item1
AND z.item2 = x.item2
AND z.max_id = x.id
WHERE x.item2 <> 0
GROUP
BY x.id;
+-------+-------+----------------------+
| item1 | item2 | COUNT(DISTINCT y.id) |
+-------+-------+----------------------+
| 1 | 2 | 2 |
| 1 | 3 | 3 |
| 2 | 3 | 1 |
+-------+-------+----------------------+