How to improve query Mysql? - mysql

Question

I made this query in MySql. It found but the time of query is 3 or + minutes.
I would like to know if is possible to improve this query.
The query is this:
SELECT CODARTIOLO, NOMEARTICOLO, SUM(QUANTITA) AS QUANTITA,
(SUM(TOTRIGA)/SUM(QUANTITA)) AS TOTALE,
(SELECT (SUM(QUANTITA * PREZZOCAD))/SUM(QUANTITA)
FROM vistacaricomagazzino cm
WHERE cm.DATA <= '$dataStart' AND cm.codarticolo=CODARTIOLO) AS PREZZOMEDIO
FROM vistascontrini c
WHERE c.DATA >= '$dataStart' AND c.DATA <= '$dataEnd'
GROUP BY NOMEARTICOLO
the table is:
VISTASCONTRINI
+--------------+--------------+------+-----+------------+-------+
| Field | Type | Null | Key | Default | Extra |
+--------------+--------------+------+-----+------------+-------+
| CODARTIOLO | varchar(13) | YES | | NULL | |
| NOMEARTICOLO | varchar(60) | YES | | NULL | |
| QUANTITA | int(11) | YES | | NULL | |
| TOTRIGA | decimal(9,2) | YES | | NULL | |
| DATA | date | NO | | 0000-00-00 | |
VISTACARICOMAGAZZINO
+-------------+---------------+------+-----+---------+-------+
| Field | Type | Null | Key | Default | Extra |
+-------------+---------------+------+-----+---------+-------+
| codordine | int(11) | NO | | 0 | |
| Quantita | int(11) | YES | | NULL | |
| PrezzoCad | decimal(10,3) | YES | | NULL | |
| codArticolo | varchar(13) | YES | | NULL | |
| Data | date | YES | | NULL | |
+-------------+---------------+------+-----+---------+-------+

Answer 1

If your tables don't have indexes (and it doesn't look like they do), then that's the first thing you need to fix. If you do that right (ie put the indexes on the fields that need to be indexed), it will probably solve the problem for you in one hit.
The fields you need to consider indexing are the ones being used by the WHERE and GROUP BY clauses.
Next, consider converting it from a nested SELECT query into a JOIN query. This will probably give you better performance too.
Finally, you haven't stated just how much data is being collated here, but if it's a large amount of data, then consider storing the collated data totals separately within the database so that you can just query it directly rather than having to re-generate all those sums and groups every time. This obviously has it's own considerations (additional storage, additional code when updating data to also update the totals, possibility of things going out of sync, etc), but if you're really suffering with performance on this, it is a valid solution.

Answer 2

I think you could try to remove the subquery from select clause. Modify the query to something like (untested, just to get the idea),
SELECT c.CODARTIOLO,c.NOMEARTICOLO,SUM(c.QUANTITA) AS QUANTITA,(SUM(c.TOTRIGA)/SUM(c.QUANTITA)) AS TOTALE,(SUM(cm.QUANTITA * cm.PREZZOCAD))/SUM(cm.QUANTITA)
FROM vistascontrini c,
vistacaricomagazzino cm
WHERE cm.codarticolo=c.CODARTIOLO
AND cm.DATA <= '$dataStart'
AND c.DATA >= '$dataStart' AND c.DATA <= '$dataEnd' group by NOMEARTICOLO
Also add indexes on DATA column of tables vistascontrini and vistacaricomagazzino .

Answer 3

Add indexes on vistascontrini.DATA, vistascontrini.NOMEARTICOLO, vistacaricomagazzino.DATA, vistacaricomagazzino.codarticolo
It's not clear do you want to group by CODARTIOLO or not and what value do you need if your group only by NOMEARTICOLO ?
Try to use following query equivalent to yours:
SELECT T.*,T1.PREZZOMEDIO FROM
(
SELECT CODARTIOLO,NOMEARTICOLO,
SUM(QUANTITA) AS QUANTITA,
(SUM(TOTRIGA)/SUM(QUANTITA)) AS TOTALE,
FROM vistascontrini c
WHERE c.DATA >= '$dataStart' AND c.DATA <= '$dataEnd'
group by NOMEARTICOLO
) AS T
LEFT JOIN
( SELECT CODARTIOLO,(SUM(QUANTITA * PREZZOCAD))/SUM(QUANTITA) as PREZZOMEDIO
FROM vistacaricomagazzino cm
WHERE cm.DATA <= '$dataStart'
GROUP BY CODARTIOLO ) as T1
ON T.CODARTIOLO=T1.CODARTIOLO