Dimensioni ridotte del database ma prestazioni molto lente. Cosa posso fare?
https://dba.stackexchange.com/questions/125085
-
29-09-2020 - |
italiano
english
français
española
中国
日本の
العربية
Deutsch
한국어
Português
RussianDomanda
Abbiamo questa installazione opencart con una dimensione del database di circa 120 MB, di cui 90 MB sono MyISAM e altri sono InnoDB.
La maggior parte delle query come:
SELECT DISTINCT *, a.*, vd.seo, p.image, p2s.price, p.product_id,
p.image, pd.name,
( SELECT price
FROM product_special ps
WHERE ps.product_id = p.product_id
AND ps.customer_group_id = '1'
AND ((ps.date_start = '0000-00-00'
OR ps.date_start < NOW())
AND (ps.date_end = '0000-00-00'
OR ps.date_end > NOW()
)
AND ps.store_id = '0'
)
ORDER BY ps.priority ASC, ps.price ASC
LIMIT 1
) AS special
FROM product_variation a, variation_description vd, product_to_category pc,
variation v, product_description pd, product_to_store_alter p2s,
product p
WHERE v.variation_group_id='18'
AND p2s.product_id = p.product_id
AND p2s.active = '1'
AND p2s.store_id = '0'
AND pc.product_id = p.product_id
AND a.variation_id = v.variation_id
AND vd.variation_id = v.variation_id
AND vd.variation_id = a.variation_id
AND v.variation_id = '1418'
AND p.product_id = pd.product_id
AND a.product_id = pd.product_id
AND p2s.quantity > '0'
AND p.product_id != '17230'
AND p.status = '1'
AND p.date_available <= NOW()
GROUP BY p.product_id
ORDER BY a.product_id DESC
LIMIT 8;
L'esecuzione della query precedente richiede circa 2 secondi.MySQL Server 5.5, key_buffer è 256 MB e innodb_buffer_pool ha 128 M. I buffer di ordinamento e join hanno 8 milioni ciascuno.
Tutto ciò in cui posso migliorare, perché tali query vengono eseguite più volte per una singola pagina.L'aiuto sarebbe apprezzato!
AGGIORNAMENTO : Spiega la dichiarazione
id select_type table type possible_keys key key_len ref rows Extra
1 PRIMARY v const PRIMARY,variation_group_id PRIMARY 8 const,const 1 Using temporary; Using filesort
1 PRIMARY vd ref PRIMARY PRIMARY 4 const 1 Using where
1 PRIMARY p ref PRIMARY,status status 1 const 13711 Using where
1 PRIMARY pd ref PRIMARY PRIMARY 4 ninec.p.product_id 1
1 PRIMARY p2s eq_ref PRIMARY PRIMARY 8 ninec.p.product_id,const 1 Using where
1 PRIMARY a ref product_id,variation_id product_id 4 ninec.p2s.product_id 14 Using where
1 PRIMARY pc ref product_id product_id 4 ninec.p.product_id 3 Using index
2 DEPENDENT SUBQUERY ps ref customer_group_id,store_id,product_id customer_group_id 4 const 1 Using where; Using filesort
AGGIORNARE
Aggiungendo CREATE TABLE affermazioni per le tabelle precedenti:
CREATE TABLE variation (
variation_id int(11) NOT NULL AUTO_INCREMENT,
variation_group_id int(11) NOT NULL,
sort_order int(3) NOT NULL,
PRIMARY KEY (variation_id),
KEY variation_group_id (variation_group_id)
) ENGINE=MyISAM
CREATE TABLE variation_description (
variation_id int(11) NOT NULL,
language_id int(11) NOT NULL,
name varchar(64) COLLATE utf8_bin NOT NULL,
description varchar(255) COLLATE utf8_bin NOT NULL,
seo varchar(255) COLLATE utf8_bin NOT NULL,
image varchar(255) COLLATE utf8_bin NOT NULL,
is_moq tinyint(1) NOT NULL,
url_link varchar(255) COLLATE utf8_bin NOT NULL,
video text COLLATE utf8_bin NOT NULL,
download_link varchar(255) COLLATE utf8_bin NOT NULL,
launch_date date NOT NULL,
PRIMARY KEY (variation_id,language_id),
KEY seo (seo),
KEY language_id (language_id)
) ENGINE=MyISAM
CREATE TABLE product (
product_id int(11) NOT NULL AUTO_INCREMENT,
model varchar(64) COLLATE utf8_bin NOT NULL,
sku varchar(64) COLLATE utf8_bin NOT NULL,
upc varchar(12) COLLATE utf8_bin NOT NULL,
location varchar(128) COLLATE utf8_bin NOT NULL,
quantity int(4) NOT NULL DEFAULT '0',
rs_quantity int(4) NOT NULL,
moq_quantity int(4) NOT NULL,
stock_status_id int(11) NOT NULL,
image varchar(255) COLLATE utf8_bin DEFAULT NULL,
watermark_pos int(1) NOT NULL,
watermark_colour int(1) NOT NULL,
manufacturer_id int(11) NOT NULL,
vendor_id varchar(25) COLLATE utf8_bin NOT NULL,
shipping tinyint(1) NOT NULL DEFAULT '1',
mrp_price decimal(15,2) NOT NULL,
price decimal(15,4) NOT NULL DEFAULT '0.0000',
org_price decimal(15,0) NOT NULL,
moq_price decimal(15,0) NOT NULL,
points int(8) NOT NULL DEFAULT '0',
tax_class_id int(11) NOT NULL,
date_available date NOT NULL,
weight decimal(15,8) NOT NULL DEFAULT '0.00000000',
weight_class_id int(11) NOT NULL DEFAULT '0',
length decimal(15,8) NOT NULL DEFAULT '0.00000000',
width decimal(15,8) NOT NULL DEFAULT '0.00000000',
height decimal(15,8) NOT NULL DEFAULT '0.00000000',
length_total float NOT NULL,
length_class_id int(11) NOT NULL DEFAULT '0',
subtract tinyint(1) NOT NULL DEFAULT '1',
minimum int(11) NOT NULL DEFAULT '1',
sort_order int(11) NOT NULL DEFAULT '0',
best_seller int(1) NOT NULL,
new_arrival int(1) NOT NULL,
readytoship int(1) NOT NULL,
exclusive int(1) NOT NULL,
sale int(1) NOT NULL,
is_readymade tinyint(1) NOT NULL,
status tinyint(1) NOT NULL DEFAULT '0',
date_added datetime NOT NULL DEFAULT '0000-00-00 00:00:00',
date_modified datetime NOT NULL DEFAULT '0000-00-00 00:00:00',
viewed int(5) NOT NULL DEFAULT '0',
time_to_ship varchar(25) COLLATE utf8_bin NOT NULL DEFAULT '5',
wash_care varchar(255) COLLATE utf8_bin NOT NULL,
measure_group_id int(11) NOT NULL,
wishlist int(11) NOT NULL,
shoppingcart int(11) NOT NULL,
barcode varchar(99) COLLATE utf8_bin NOT NULL,
PRIMARY KEY (product_id),
KEY sort_order (sort_order),
KEY best_seller (best_seller),
KEY new_arrival (new_arrival),
KEY readytoship (readytoship),
KEY exclusive (exclusive),
KEY sale (sale),
KEY status (status),
KEY viewed (viewed),
KEY time_to_ship (time_to_ship),
KEY model (model)
) ENGINE=InnoDB
CREATE TABLE product_description (
product_id int(11) NOT NULL AUTO_INCREMENT,
language_id int(11) NOT NULL,
name varchar(255) COLLATE utf8_bin NOT NULL,
description text COLLATE utf8_bin NOT NULL,
meta_description varchar(255) COLLATE utf8_bin NOT NULL,
meta_keyword varchar(255) COLLATE utf8_bin NOT NULL,
PRIMARY KEY (product_id,language_id),
KEY name (name),
KEY language_id (language_id)
) ENGINE=MyISAM
CREATE TABLE product_special (
product_special_id int(11) NOT NULL AUTO_INCREMENT,
product_id int(11) NOT NULL,
customer_group_id int(11) NOT NULL,
priority int(5) NOT NULL DEFAULT '1',
price decimal(15,4) NOT NULL DEFAULT '0.0000',
date_start date NOT NULL DEFAULT '0000-00-00',
date_end date NOT NULL DEFAULT '0000-00-00',
store_id int(11) NOT NULL,
PRIMARY KEY (product_special_id),
KEY product_id (product_id),
KEY customer_group_id (customer_group_id),
KEY store_id (store_id)
) ENGINE=MyISAM
CREATE TABLE product_to_category (
product_id int(11) NOT NULL,
category_id int(11) NOT NULL,
KEY product_id (product_id,category_id),
KEY category_id (category_id,product_id)
) ENGINE=InnoDB
CREATE TABLE product_to_store_alter (
product_id int(11) NOT NULL,
store_id int(11) NOT NULL DEFAULT '0',
location varchar(128) NOT NULL,
quantity int(10) NOT NULL DEFAULT '0',
price decimal(15,4) NOT NULL DEFAULT '0.0000',
substract tinyint(1) NOT NULL DEFAULT '1',
minimum int(11) NOT NULL DEFAULT '1',
active tinyint(4) NOT NULL DEFAULT '0',
PRIMARY KEY (product_id,store_id)
) ENGINE=MyISAM
CREATE TABLE product_variation (
product_id int(11) NOT NULL,
variation_id varchar(255) COLLATE utf8_bin NOT NULL,
language_id int(11) NOT NULL,
KEY product_id (product_id),
KEY variation_id (variation_id)
) ENGINE=MyISAM
Soluzione
Usa la sintassi JOIN ... ON ..., per favore.
INDEX(customer_group_id, product_id, store_id) sarà un indice migliore per ps.
Non introdurre la sottoquery dipendente finché non hai un elenco DISTINCT di product_ids. In questo modo non lo valuterai ripetutamente.Questo è,
- rimuovere la sottoquery ("speciale")
usa la query rimanente come sottoquery:
SELECT *, ( SELECT price ... ) AS special FROM ( SELECT DISTINCT ... ORDER BY ... LIMIT 8 ) AS x;
Ora eseguirai la sottoquery solo 8 volte, non più di un milione di volte.
Probabilmente altre tabelle potrebbero essere ritardate in questo modo;Non ho la pazienza di guadare quella tabella 7 JOIN. Probabilmente pc e a.
Sii coerente su quale product_id stai usando nel GROUP BY e nel ORDER BY;sembrano essere tutti uguali e probabilmente l'ottimizzatore è abbastanza intelligente.Suggerisci di usare p. poiché è il primo nel EXPLAIN.
Fai uno dei KEYs nel product_to_category il PRIMARY KEY. (Suppongo che tu non voglia duplicati?)
Non preoccuparti di indicizzare le bandiere (vendita, viste, ...)
