Optimizing MySQL query with INNER JOINS, LEFT JOINS, GROUP BY and HAVING
Question
I'm having trouble optimizing this really big query and I can't change the table structure except creating additional indexes and small adjustments.
SELECT
'Fattura Prodotti Postali' AS `type`,
SUM(dpd.qta) AS `products_count_quantity`,
COUNT(dpd.IDlavorazione_dett) AS `products_count`,
GROUP_CONCAT(DISTINCT dp.prod_totali - CAST(dp.opzione1 AS UNSIGNED) SEPARATOR ' |-| ') AS `process_products_count`,
GROUP_CONCAT(DISTINCT dp.IDdistinta) AS `product_code`,
GROUP_CONCAT(DISTINCT dp.data_distinta) AS `process_date`,
GROUP_CONCAT(DISTINCT dp.IDesito) AS `process_status_id`,
GROUP_CONCAT(DISTINCT dp.note) AS `process_note`,
GROUP_CONCAT(DISTINCT dp.IDlavorazione) AS `unique_id`,
SUM(dpd.tariffa) AS `products_total`,
SUM(IF(o.IDdoc IS NULL,1,0)*dpd.qta) AS `tobill_count_quantity`,
SUM(IF(NOT o.IDdoc IS NULL,1,0)*dpd.qta) AS `billed_count_quantity`,
SUM(IF(o.IDdoc IS NULL,1,0)) AS `tobill_count`,
SUM(IF(o.IDdoc IS NULL, dpd.tariffa,0)) AS `tobill_total`,
SUM(IF(NOT o.IDdoc IS NULL,1,0)) AS `billed_count`,
SUM(IF(NOT o.IDdoc IS NULL, dpd.tariffa,0)) AS `billed_total`,
SUM(IF(o.IDdoc IS NULL, dpd.tariffa,0)-IF(NOT o.IDdoc IS NULL, dpd.tariffa,0)) AS `bill_diff`,
COUNT(dpd.IDlavorazione_dett) AS `products_count`,
SUM(dpd.tariffa) AS `products_total`,
SUM(dpd.tariffa*(dpd.iva/100)) AS `products_vat`,
SUM(dpd.tariffa*(dpd.sconto/100.0)) AS `products_discount`,
SUM(dpd.tariffa*(1.0-dpd.sconto/100.0)*(dpd.iva/100)) AS `products_discount_vat`,
SUM(dpd.tariffa*(1+(dpd.iva/100.0))) AS `products_total_vat`,
SUM(ROUND(dpd.tariffa*(1.0-dpd.sconto/100.0),5)) AS `products_total_discount`,
SUM(dpd.tariffa*(1.0-dpd.sconto/100.0)*(1+(dpd.iva/100.0))) AS `products_total_discount_vat`,
SUM(dpd.qta) AS `products_quantity`,
SUM(IF(o.IDdoc IS NULL,1,0)) AS `tobill_count`,
SUM(IF(o.IDdoc IS NULL, dpd.tariffa,0)) AS `tobill_total`,
SUM(IF(o.IDdoc IS NULL, dpd.tariffa*(1+(dpd.iva/100.0)),0)) AS `tobill_total_vat`,
SUM(IF(NOT o.IDdoc IS NULL,1,0)) AS `billed_count`,
SUM(IF(NOT o.IDdoc IS NULL, dpd.tariffa,0)) AS `billed_total`,
SUM(IF(NOT o.IDdoc IS NULL, dpd.tariffa*(1+(dpd.iva/100.0)),0)) AS `billed_total_vat`
FROM doc_prodottipostali_dett dpd
INNER JOIN tracking t ON (dpd.IDlavorazione_dett=t.product_id)
LEFT JOIN prodotti_pp ppp ON (dpd.IDprodotto=ppp.IDprodotto)
LEFT JOIN categorie_pp cpp ON (ppp.categoria=cpp.IDcategoria)
LEFT JOIN categorie_pp cppp ON (CAST(dpd.IDcategoria AS UNSIGNED)=cppp.IDcategoria)
INNER JOIN doc_prodottipostali dp ON (dpd.IDlavorazione=dp.IDlavorazione)
LEFT JOIN ordini o ON (dpd.IDfattura=o.IDdoc)
WHERE
(
(dp.tipo = 'PT' AND t.date >= '2022-05-15 00:00:00' AND t.date <= '2022-07-26 23:59:59') AND
((IF(dpd.IDprodotto>0, cpp.codice, IF(CAST(dpd.IDcategoria AS UNSIGNED)>0, cppp.codice, 'NONE')) NOT IN ('LAW','AR','CAD','EMESSOCAD') OR IF(dpd.IDprodotto>0, cpp.codice, IF(CAST(dpd.IDcategoria AS UNSIGNED)>0, cppp.codice, 'NONE')) IS NULL)) AND
((t.last IN (-1,2,3,6,7,10,11,34,35,130,131,258,514,4098,4354,8194,8450))) AND
(((dp.opzione2 = 'PI' AND dp.data_distinta < '2022-08-25')) OR ((dp.data_distinta >= '2022-08-25')) OR ((o.IDdoc >= '1'))) AND
(((t.last IN (-1,2,3,6,7,10,11,34,35,130,131,258,514,4098,4354,8194,8450))))
) AND
(
((NOT t.last IN (-1,4,5,6,7,20,21,132,133,149,516,532,1028,1157,8197)))
)
GROUP BY dpd.IDlavorazione
HAVING (1=1 AND ((tobill_count > '0')))
ORDER BY dp.data_distinta ASC;
The create tables are as following:
CREATE TABLE `doc_prodottipostali_dett` (
`IDlavorazione_dett` int(11) NOT NULL AUTO_INCREMENT,
`IDlavorazione` int(11) NOT NULL,
`IDdistinta` varchar(45) NOT NULL,
`IDcategoria` varchar(45) NOT NULL,
`IDprodotto` int(11) NOT NULL,
`codiceabarre` varchar(255) NOT NULL,
`codiceavviso` varchar(255) NOT NULL DEFAULT '',
`IDvettore` int(11) NOT NULL,
`IDlistino` int(11) NOT NULL,
`rif` varchar(45) NOT NULL,
`IDmittente` int(11) NOT NULL,
`IDdestinatario` int(11) NOT NULL,
`ufficio_mittente` varchar(45) NOT NULL,
`nome_lavoro` varchar(255) NOT NULL,
`IDpostino` int(11) DEFAULT NULL,
`note` longtext NOT NULL,
`allegati` char(1) NOT NULL,
`utente` varchar(45) NOT NULL,
`peso` decimal(10,5) NOT NULL,
`tariffa` decimal(10,5) NOT NULL,
`sconto` float NOT NULL DEFAULT 0,
`iva` int(11) NOT NULL,
`qta` int(11) NOT NULL,
`am` char(1) NOT NULL,
`cp` char(1) NOT NULL,
`eu` char(1) NOT NULL,
`aa` char(10) NOT NULL,
`ee` char(10) NOT NULL,
`stato` int(11) NOT NULL,
`lavorato` int(1) NOT NULL,
`IDesito` int(11) NOT NULL,
`esito` varchar(45) NOT NULL,
`data_op` datetime NOT NULL,
`fatturato` int(11) NOT NULL,
`data_fatt` date NOT NULL,
`IDfattura` int(11) NOT NULL,
`data` datetime DEFAULT NULL,
`data_ar` datetime DEFAULT NULL,
`nome_ar` varchar(255) DEFAULT NULL,
`file_ar` varchar(255) DEFAULT NULL,
`IDflusso_dett` int(11) NOT NULL DEFAULT 0,
`type` varchar(2) NOT NULL DEFAULT '',
`typology` varchar(2) NOT NULL DEFAULT '',
`related_id` int(11) NOT NULL DEFAULT 0,
`related_ar_id` int(11) NOT NULL DEFAULT 0,
`unregistered` int(1) NOT NULL DEFAULT 0,
`total_attachments` int(11) NOT NULL DEFAULT 0,
`repeated_recipient` int(4) DEFAULT NULL,
`law_tomanage` tinyint(4) NOT NULL DEFAULT 0,
`law_towork` int(11) NOT NULL DEFAULT 0,
`law_toprint` int(11) NOT NULL DEFAULT 0,
`notlaw_tocomplete` int(11) NOT NULL DEFAULT 0,
PRIMARY KEY (`IDlavorazione_dett`),
KEY `IDvettore` (`IDvettore`),
KEY `IDmittente` (`IDmittente`),
KEY `IDdestinatario` (`IDdestinatario`),
KEY `IDlavorazione` (`IDlavorazione`),
KEY `IDfattura` (`IDfattura`),
KEY `codiceabarre` (`codiceabarre`),
KEY `rif` (`rif`),
KEY `IDpostino` (`IDpostino`),
KEY `IDcategoria` (`IDcategoria`),
KEY `IDprodotto` (`IDprodotto`),
KEY `IDlistino` (`IDlistino`),
KEY `codiceavviso` (`codiceavviso`),
KEY `IDflusso_dett` (`IDflusso_dett`),
KEY `typology` (`typology`),
KEY `nome_ar` (`nome_ar`),
KEY `unregistered` (`unregistered`),
KEY `related_ar_id` (`related_ar_id`),
KEY `related_id` (`related_id`) USING BTREE,
KEY `type` (`type`),
KEY `stato` (`stato`),
KEY `lavorato` (`lavorato`),
KEY `law_tomanage` (`law_tomanage`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8;
CREATE TABLE `tracking` (
`tracking_id` int(11) NOT NULL AUTO_INCREMENT,
`product_id` int(11) NOT NULL,
`status_id` int(11) NOT NULL,
`lat` double NOT NULL DEFAULT 0,
`lng` double NOT NULL DEFAULT 0,
`entity_id` int(11) NOT NULL,
`sub_entity_id` int(11) NOT NULL DEFAULT 0,
`date` datetime NOT NULL,
`note` text DEFAULT NULL,
`last` int(11) NOT NULL DEFAULT 0,
`date_last` datetime NOT NULL DEFAULT current_timestamp(),
`tracking_rel` int(11) NOT NULL DEFAULT 0,
`quantity_from` int(11) NOT NULL DEFAULT 1,
`quantity_to` int(11) NOT NULL DEFAULT 1,
`price` decimal(14,4) NOT NULL DEFAULT 0.0000,
`price_unit` decimal(14,4) NOT NULL DEFAULT 0.0000,
`vat` decimal(14,4) NOT NULL DEFAULT 0.0000,
`invoice_id` int(11) NOT NULL DEFAULT 0,
`management_status` int(11) NOT NULL DEFAULT 0,
`package_id` int(11) NOT NULL DEFAULT 0,
PRIMARY KEY (`tracking_id`),
KEY `product_id` (`product_id`),
KEY `status_id` (`status_id`),
KEY `entity_id` (`entity_id`),
KEY `sub_entity_id` (`sub_entity_id`),
KEY `date` (`date`),
KEY `last` (`last`),
KEY `package_id` (`package_id`),
KEY `management_status` (`management_status`),
KEY `date_last` (`date_last`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8;
CREATE TABLE `prodotti_pp` (
`IDprodotto` int(10) unsigned NOT NULL AUTO_INCREMENT,
`codice` varchar(255) NOT NULL DEFAULT '',
`codiceabarre` varchar(255) NOT NULL DEFAULT '',
`descrizione` longtext NOT NULL,
`prezzo` decimal(10,2) NOT NULL DEFAULT 0.00,
`prezzo_2` decimal(10,2) NOT NULL DEFAULT 0.00,
`prezzo_3` decimal(10,2) NOT NULL DEFAULT 0.00,
`UM` char(3) NOT NULL DEFAULT '',
`peso` decimal(10,2) NOT NULL DEFAULT 0.00,
`datapv` date NOT NULL DEFAULT '0000-00-00',
`iva` decimal(10,0) NOT NULL DEFAULT 0,
`fornitore` varchar(45) NOT NULL DEFAULT '',
`categoria` int(10) unsigned NOT NULL DEFAULT 0,
`trasp` decimal(10,2) NOT NULL DEFAULT 0.00,
`dettaglio` longtext NOT NULL,
`ico` varchar(255) NOT NULL DEFAULT '',
`foto` varchar(255) NOT NULL DEFAULT '',
`visibility_web` char(1) NOT NULL DEFAULT '',
`composito` char(1) NOT NULL DEFAULT '',
`sottocat` int(10) unsigned NOT NULL DEFAULT 0,
`tipologia` int(10) NOT NULL DEFAULT 0,
`marchio` int(10) unsigned NOT NULL DEFAULT 0,
`disp` char(1) NOT NULL DEFAULT '',
`vetrina1` char(1) NOT NULL DEFAULT '',
`vetrina2` char(1) NOT NULL DEFAULT '',
`click` decimal(10,0) NOT NULL DEFAULT 0,
`IDnote` int(11) NOT NULL DEFAULT 1,
PRIMARY KEY (`IDprodotto`),
KEY `categoria` (`categoria`),
KEY `codice` (`codice`)
) ENGINE=InnoDB DEFAULT CHARSET=UTF8;
CREATE TABLE `categorie_pp` (
`IDcategoria` int(10) unsigned NOT NULL AUTO_INCREMENT,
`IDmadre` int(11) NOT NULL DEFAULT 0,
`codice` varchar(255) NOT NULL,
`nome` varchar(45) NOT NULL DEFAULT '',
`visibility_web` char(1) NOT NULL DEFAULT '',
PRIMARY KEY (`IDcategoria`),
KEY `codice` (`codice`),
KEY `IDmadre` (`IDmadre`)
) ENGINE=InnoDB DEFAULT CHARSET=UTF8;
CREATE TABLE `doc_prodottipostali` (
`IDlavorazione` int(11) NOT NULL AUTO_INCREMENT,
`codice_lavorazione` varchar(255) NOT NULL,
`IDdistinta` varchar(45) NOT NULL,
`tipo` varchar(45) NOT NULL,
`data_distinta` date NOT NULL,
`data_lavorazione` date NOT NULL,
`IDcliente` int(11) DEFAULT 1,
`IDpagamento` int(11) NOT NULL,
`note` longtext NOT NULL,
`utente` varchar(45) NOT NULL,
`stato` int(11) NOT NULL,
`data_op` datetime NOT NULL,
`opzioni` varchar(45) NOT NULL,
`rif` varchar(45) NOT NULL,
`opzione1` varchar(45) NOT NULL,
`opzione2` varchar(45) NOT NULL,
`allegati` int(11) NOT NULL DEFAULT 0,
`IDesito` int(11) NOT NULL,
`esito` varchar(45) NOT NULL,
`prod_totali` int(11) NOT NULL,
`prod_accettati` int(11) NOT NULL,
`prod_fatturati` int(11) NOT NULL,
`prod_chiusi` int(11) NOT NULL,
`prod_end_shipping` int(11) NOT NULL,
`stato_fatt` char(1) NOT NULL,
`IDfattura` int(11) NOT NULL DEFAULT 0,
`IDrel` int(11) NOT NULL DEFAULT 0,
`IDflow` int(11) NOT NULL DEFAULT 0,
`laws_tomanage` tinyint(4) NOT NULL DEFAULT 0,
`idSender` int(11) NOT NULL,
PRIMARY KEY (`IDlavorazione`),
KEY `IDdistinta` (`IDdistinta`),
KEY `IDfattura` (`IDfattura`),
KEY `IDrel` (`IDrel`),
KEY `allegati` (`allegati`),
KEY `tipo` (`tipo`),
KEY `IDesito` (`IDesito`),
KEY `laws_tomanage` (`laws_tomanage`),
KEY `data_distinta` (`data_distinta`)
) ENGINE=InnoDB DEFAULT CHARSET=UTF8;
CREATE TABLE `ordini` (
`IDdoc` int(11) NOT NULL AUTO_INCREMENT,
`IDsoggetto` int(11) DEFAULT 1,
`numero` varchar(45) NOT NULL DEFAULT '0',
`numero_web` int(11) NOT NULL,
`datadoc` date NOT NULL DEFAULT '0000-00-00',
`tipodoc` varchar(255) NOT NULL,
`type_doc` varchar(4) NOT NULL DEFAULT '',
`modpag` int(10) unsigned NOT NULL DEFAULT 0,
`div_dest` longtext NOT NULL,
`IDagente` varchar(6) NOT NULL DEFAULT '0',
`IDvettore` int(10) unsigned NOT NULL DEFAULT 0,
`imballo` decimal(20,5) NOT NULL DEFAULT 0.00000,
`colli` decimal(10,0) NOT NULL DEFAULT 0,
`datascad` date NOT NULL DEFAULT '0000-00-00',
`stato` char(1) NOT NULL DEFAULT '',
`note` longtext NOT NULL,
`importo` decimal(20,5) NOT NULL DEFAULT 0.00000,
`tipo` varchar(4) NOT NULL DEFAULT '',
`causale` varchar(45) NOT NULL DEFAULT '',
`azione` char(1) NOT NULL DEFAULT '',
`utente` varchar(45) NOT NULL DEFAULT '',
`data_op` datetime NOT NULL DEFAULT '0000-00-00 00:00:00',
`acconto` decimal(20,5) NOT NULL DEFAULT 0.00000,
`iva` decimal(20,5) NOT NULL DEFAULT 0.00000,
`sconto` decimal(20,5) NOT NULL DEFAULT 0.00000,
`n_doc_passive` varchar(45) NOT NULL DEFAULT '',
`porto` varchar(45) NOT NULL DEFAULT '',
`dataora_rit` datetime NOT NULL DEFAULT '0000-00-00 00:00:00',
`ordine_web` char(1) NOT NULL DEFAULT '',
`clonato` char(1) NOT NULL DEFAULT '',
`trasp` decimal(20,5) NOT NULL DEFAULT 0.00000,
`varie` decimal(20,5) NOT NULL DEFAULT 0.00000,
`banca_appo` int(11) NOT NULL DEFAULT 0,
`IDmagazzino` int(11) NOT NULL DEFAULT 0,
`pv` int(11) NOT NULL DEFAULT 0,
`fornitore` int(11) NOT NULL DEFAULT 0,
`qta_colli` decimal(10,0) NOT NULL DEFAULT 0,
`bolli` decimal(20,5) NOT NULL DEFAULT 0.00000,
`agibilitamezzi` varchar(255) NOT NULL DEFAULT '',
`opzioneprezzo` varchar(45) NOT NULL DEFAULT '',
`emailcc` varchar(255) NOT NULL DEFAULT '',
`U_IBAN` varchar(45) NOT NULL,
`stato_fatt` char(1) NOT NULL,
`IDfattura` int(11) NOT NULL,
`data_fatt` date NOT NULL,
`digital_idinvoice` varchar(255) NOT NULL DEFAULT '0',
`digital_idupload` bigint(11) NOT NULL DEFAULT 0,
`digital_status` int(11) NOT NULL DEFAULT 0,
`digital_type` varchar(2) NOT NULL DEFAULT '',
`digital_office` int(11) NOT NULL,
`digital_sectional` char(2) NOT NULL DEFAULT '',
`digital_last_office` int(11) NOT NULL,
`digital_flag` varchar(4) NOT NULL,
`determines_code` varchar(255) NOT NULL,
`determines_date` date NOT NULL,
`determines_id` varchar(255) NOT NULL,
`flag_2` varchar(255) NOT NULL,
`flag` int(11) NOT NULL,
`flag_note` longtext NOT NULL,
`flag_3` varchar(255) NOT NULL,
`type_op` varchar(255) NOT NULL,
`IDpartenza_rel` int(11) NOT NULL,
`IDdestinazione_rel` int(11) NOT NULL,
`doc_insinuazionepassivo` varchar(255) NOT NULL,
`doc_insinuazionepassivo_data` date NOT NULL DEFAULT '0000-00-00',
PRIMARY KEY (`IDdoc`),
KEY `IDsoggetto` (`IDsoggetto`),
KEY `IDfattura` (`IDfattura`),
KEY `pv` (`pv`),
KEY `azione` (`azione`),
KEY `numero` (`numero`),
KEY `tipodoc` (`tipodoc`),
KEY `datadoc` (`datadoc`),
KEY `digital_sectional` (`digital_sectional`),
KEY `digital_type` (`digital_type`),
KEY `digital_office` (`digital_office`),
KEY `digital_status` (`digital_status`),
KEY `type_doc` (`type_doc`),
KEY `flag_2` (`flag_2`),
KEY `flag` (`flag`),
KEY `flag_3` (`flag_3`),
KEY `type_op` (`type_op`)
) ENGINE=InnoDB DEFAULT CHARSET=UTF8;
These are the number of rows for each table:
tracking:42231628
doc_prodottipostali_det:11316150
doc_prodottipostali:40556
ordini:40360
prodotti_pp:52
categorie_pp:30
I've tried making a subquery on the tracking table (the biggest), trying to filter out most of the rows but it didn't work.
This is the explain I get, it looks fine except for the "Using temporary" and "Using filesort".
The problem with the query is that with low numbers of elaborated rows it runs fine, the moment it tries to elaborate more than 1 million it begins to slow down.
Explain
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
|---|---|---|---|---|---|---|---|---|---|
| 1 | SIMPLE | t | range | product_id,date,last | date | 5 | \N | 69266 | Using index condition; Using where; Using temporary; Using filesort |
| 1 | SIMPLE | dpd | eq_ref | PRIMARY,IDlavorazione | PRIMARY | 4 | db.t.product_id | 1 | |
| 1 | SIMPLE | ppp | eq_ref | PRIMARY | PRIMARY | 4 | db.dpd.IDprodotto | 1 | Using where |
| 1 | SIMPLE | cpp | eq_ref | PRIMARY | PRIMARY | 4 | db.ppp.categoria | 1 | Using where |
| 1 | SIMPLE | cppp | eq_ref | PRIMARY | PRIMARY | 4 | func | 1 | Using where |
| 1 | SIMPLE | o | eq_ref | PRIMARY | PRIMARY | 4 | db.dpd.IDfattura | 1 | Using index |
| 1 | SIMPLE | dp | eq_ref | PRIMARY,tipo,data_distinta | PRIMARY | 4 | db.dpd.IDlavorazione | 1 | Using where |
When you try to extend the time range on the tracking table t.date >= '2022-05-15 00:00:00' AND t.date <= '2022-07-26 23:59:59' the optimizer switches to a full table scan. I've tried forcing the 'date' index on the tracking table but it still slows down with big ranges. The users are supposed to search without the time range too, so an index on the date isn't really the best option.
Update: I've fixed the malformed query by aggregating the nonggregated fields. The GROUP_CONCAT fields returned the same values and the ONLY_FULL_GROUP_BY setting was disabled, that's why it was working normally.
I'll be more specific on the problems that I'm facing. When a user tries to search with a more extended range of dates, for example from '2022-03-15 00:00:00' to '2022-07-26 23:59:59' the query takes 5:24 minutes to complete with the following explain:
Explain with extended date range:
| id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra |
|---|---|---|---|---|---|---|---|---|---|
| 1 | SIMPLE | dp | ref | PRIMARY,tipo,data_distinta | tipo | 137 | const | 20400 | Using index condition; Using where; Using temporary; Using filesort |
| 1 | SIMPLE | dpd | ref | PRIMARY,IDlavorazione | IDlavorazione | 4 | db.dp.IDlavorazione | 115 | |
| 1 | SIMPLE | ppp | eq_ref | PRIMARY | PRIMARY | 4 | db.dpd.IDprodotto | 1 | Using where |
| 1 | SIMPLE | cpp | eq_ref | PRIMARY | PRIMARY | 4 | db.ppp.categoria | 1 | Using where |
| 1 | SIMPLE | cppp | eq_ref | PRIMARY | PRIMARY | 4 | func | 1 | Using where |
| 1 | SIMPLE | o | eq_ref | PRIMARY | PRIMARY | 4 | db.dpd.IDfattura | 1 | Using where; Using index |
| 1 | SIMPLE | t | ref | product_id,date,last | product_id | 4 | db.dpd.IDlavorazione_dett | 1 | Using where |
Update 2: Removing the ORDER BY clause removes the 'Using temporary' and 'Using filesort' from the explain, I guess it's an index issue, even if the fields are covered by indexes. After removing the ORDER BY clause the query time is still 5 minutes.
2 answers
solution1
0 2022-07-29 12:11:48
Your query appears to filter your tracking table on IN-lists on last and a date range on date . A multi-column index on those two columns, with the equality-matched column first and the range-matched column second may help you. And your query uses product_id from that table. So this covering index is worth a try.
ALTER TABLE tracking
ADD INDEX last_date (last, date, product_id);
You can look up covering indexes to learn more.
solution2
0 2022-07-30 01:22:08
For starters, give these indexes a try:
dpd: INDEX(IDprodotto, IDcategoria, IDlavorazione_dett, IDlavorazione, IDfattura)
dpd: INDEX(IDlavorazione_dett, IDlavorazione, IDprodotto, IDcategoria, IDfattura)
dp: INDEX(tipo, opzione2, data_distinta, IDlavorazione)
dp: INDEX(IDlavorazione)
o: INDEX(IDdoc)
t: INDEX(date, last, product_id)
t: INDEX(product_id)
ppp: INDEX(IDprodotto, categoria)
categorie_pp: INDEX(IDcategoria)
categorie_pp: INDEX(codice, IDcategoria)
When adding a composite index, DROP any plain index(es) with the same leading columns. That is, when you have both INDEX(a) and INDEX(a,b), toss the former.
CAST(dpd.IDcategoria AS UNSIGNED) -- Perhaps you should change the datatype of IDcategoria so you can avoid the conversion here? (That might lead to better index usage.)
The main things slowing down the query ( and making it difficult to optimize):
- WHERE clause referencing multiple tables.
- OR
- NOT
- GROUP BY after JOINing 7 tables
But I don't see how to improve on any of them.
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.
Related Question
optimizing a query using union left joins and inner joins
Optimizing MySQL query with multiple left joins
Optimizing MySQL "OR" query in joins
HAVING, JOINS, GROUP BY MySQL
MYSQL query with 3 INNER JOINs
mysql inner joins query
2 LEFT JOINs in a MySQL Query
MySQL: optimizing a query with three joins
Optimizing a MySQL query with several joins
Optimizing MySQL query with several JOINs
Related Tags