Question

I have 2 tables:

stock:

pid name qty
--- ---- ---
1   aaaa   2
2   bbbb   3
1   aaaa   5
3   cccc   1
2   bbbb   2

stock_total:

pid name total_qty
--- ---- ---------

I can insert rows from stock table with the total qty to stock_total using this query

INSERT INTO stock_total (pid, name, total_qty)
SELECT pid, name, SUM(qty) 
FROM stock
GROUP BY pid, name

The problem is, I will run the SQL above via cron job. So on the next execution, the SQL should UPDATE existing product and INSERT non-exist products.

It would be very inefficient if I loop over the SELECT results, check each row if exists in stock_total and do the INSERT or UPDATE.

Is there any simpler way for achieving this? perhaps by modifying the SQL above. Thanks.

Answer 1

After some googling and experimenting with the answers here, I came up with this solution. It turns out that MySQL supports REPLACE INTO... and ...ON DUPLICATE KEY UPDATE. So my query would be like this:

REPLACE INTO stock_total
SELECT pid, name, SUM(qty)
FROM stock
GROUP by pid, name

or,

INSERT INTO stock_total
SELECT pid, name, SUM(qty)
FROM stock
GROUP by pid, name
ON DUPLICATE KEY UPDATE total_qty=VALUES(total_qty)

If a row exists on stock_total, the first query will DELETE then INSERT the new row,
and the second query will UPDATE the existing row.

Both query will only work if the table has a primary key or unique index:

CREATE TABLE stock_total (
    pid         INT         NOT NULL,
    name        VARCHAR(20) NOT NULL,
    total_qty   INT         NOT NULL,

    UNIQUE (pid, name)
);

Documentation:

• REPLACE syntax
• INSERT … ON DUPLICATE KEY UPDATE syntax