I have table name builds it look likes
CREATE TABLE `builds` (
`id` int(10) unsigned NOT NULL AUTO_INCREMENT,
`testplan_id` int(10) unsigned NOT NULL DEFAULT '0',
`name` varchar(100) NOT NULL DEFAULT 'undefined',
`creation_ts` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP,
PRIMARY KEY (`id`),
UNIQUE KEY `name` (`testplan_id`,`name`),
KEY `testplan_id` (`testplan_id`)
) ENGINE=InnoDB AUTO_INCREMENT=2074288 DEFAULT CHARSET=utf8'
I want to create partitions on build based on creation_ts. I am trying
Alter Table builds PARTITION BY RANGE (TO_DAYS(creation_ts))
( PARTITION p1 values less than (TO_DAYS('2015-05-07'))
in phpMyadmin but it is showing error "unrecognised alter operation";
I am using MYSQL Server version: 5.7.11.
Related
I have an old database with version 5.0.95 (which unfortunately I cannot upgrade).
In this database I have a table clientes_states:
CREATE TABLE `clientes_states` (
`id` int(11) NOT NULL auto_increment,
`client_id` int(11) NOT NULL,
`state_id` int(11) NOT NULL,
`date` datetime NOT NULL,
PRIMARY KEY (`id`)
) ENGINE=MyISAM AUTO_INCREMENT=1239670 DEFAULT CHARSET=latin1 ROW_FORMAT=FIXED;
For each change of a client state, a new entry is generated. From this reason this table is huge, which makes it complicated to achieve last state of all clients.
In order to access easily the current state of each client, I created the following table, and added a trigger:
CREATE TABLE `actual_clients_states` (
`id` int(11) NOT NULL auto_increment,
`client_id` int(11) NOT NULL,
`state_id` int(11) NOT NULL,
`updated_at` timestamp NOT NULL default CURRENT_TIMESTAMP,
PRIMARY KEY (`id`),
UNIQUE KEY `client_id` (`client_id`)
) ENGINE=InnoDB AUTO_INCREMENT=21022 DEFAULT CHARSET=utf8 ROW_FORMAT=COMPACT;
CREATE TRIGGER clients_states_insert
AFTER INSERT
ON clientes_states FOR EACH ROW
INSERT INTO actual_clients_states (client_id, state_id)
VALUES (new.client_id, new.state_id)
ON DUPLICATE KEY UPDATE state_id=NEW.state_id;
I have two databases, one for testing and one for staging.
From some reason, the updated_at field is updated with wrong date (almost a month ago) on the staging server, but works fine on the testing server.
Any idea what might be the problem?
I have implemented custom url shortener in my app and I have one table for that. table structure looks like this:
CREATE TABLE `urls` (
`id` int(11) NOT NULL,
`url_id` varchar(10) DEFAULT NULL,
`long_url` varchar(255) DEFAULT NULL,
`clicked` mediumint(5) NOT NULL DEFAULT 0,
`user_id` varchar(7) DEFAULT NULL,
`type` varchar(15) DEFAULT NULL,
`ad_id` int(11) DEFAULT NULL,
`campaign` int(11) DEFAULT,
`increment` tinyint(1) NOT NULL DEFAULT 0,
`date` date DEFAULT NULL,
`del` enum('1','0') NOT NULL DEFAULT '0'
) ENGINE=InnoDB DEFAULT CHARSET=utf8 ROW_FORMAT=COMPACT
ALTER TABLE `urls`
ADD PRIMARY KEY (`id`),
ADD KEY `url_id` (`url_id`),
ADD KEY `type` (`type`),
ADD KEY `campaign` (`campaign`),
ADD KEY `ad_id` (`ad_id`),
ADD KEY `date` (`date`),
ADD KEY `user_id` (`user_id`);
The table now has 20.000.000 records and currently growing by 300k-400k records per day.
url_id column is unique varchar(10) and url looks like that: http://example.com/asdfghjklu
Now i have partitioned this table into 10 partitions by HASH(id):
PARTITION BY HASH (`id`)
PARTITIONS 10;
When I try to generate reports and join this table on others query is getting really slow, so slow even can't get 1 week report.
When I try to make big query in this table I filter almost every query with dates and I think it will be much better if I partition this table by date column.
Is it good idea?
As I read if I want to partition this table by date I need to add date in composite primary key: PRIMARY KEY(id, date)
What do you think about this? How do I improve my query performance?
I wold recommend use hash partition using date or month or YEAR
CREATE TABLE `urls` (
`id` int(11) NOT NULL,
`url_id` varchar(10) DEFAULT NULL,
`long_url` varchar(255) DEFAULT NULL,
`clicked` mediumint(5) NOT NULL DEFAULT 0,
`user_id` varchar(7) DEFAULT NULL,
`type` varchar(15) DEFAULT NULL,
`ad_id` int(11) DEFAULT NULL,
`campaign` int(11) DEFAULT,
`increment` tinyint(1) NOT NULL DEFAULT 0,
`date` date DEFAULT NULL,
`del` enum('1','0') NOT NULL DEFAULT '0',
PartitionsID int(4) unsigned NOT NULL,
KEY PartitionsID (PartitionsID)
) ENGINE=InnoDB DEFAULT CHARSET=latin1
PARTITION BY HASH (PartitionsID)
PARTITIONS 366;
IN PARTITION ID you just need to insert TO_DAYS(date) so you have only one value for entire day .
SOURCE
and it will make easy for partition for each day or you can do with month wise also depending on your data size .
for select
you can use below query as example
SELECT *
FROM TT ACT
WHERE ACT.CustomerID = vCustomerID
AND ACT.TransactionTime BETWEEN vInvoiceEndDate AND vPaymentDueDate
AND ACT.TrxnInfoTypeID IN (19, 23)
AND ACT.PaymentType = '1'
AND ACT.PartitionsID BETWEEN TO_DAYS(vInvoiceEndDate) AND TO_DAYS(vPaymentDueDate);
I have a table with the following scheme :
CREATE TABLE `type_interests` (
`id` int(11) NOT NULL,
`interest_id` int(11) NOT NULL,
`type_id` int(11) NOT NULL,
`type` varchar(64) NOT NULL,
`status` varchar(64) NOT NULL,
`created_date` datetime NOT NULL,
`last_update` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
`notes` varchar(64) NOT NULL
) ENGINE=InnoDB DEFAULT CHARSET=utf8;
ALTER TABLE `type_interests`
ADD PRIMARY KEY (`id`),
ADD KEY `interest_id` (`interest_id`,`type_id`),
ADD KEY `interest_id_2` (`interest_id`);
ALTER TABLE `type_interests`
MODIFY `id` int(11) NOT NULL AUTO_INCREMENT, AUTO_INCREMENT=31415;
ALTER TABLE `type_interests`
ADD CONSTRAINT `type_interests_ibfk_1` FOREIGN KEY (`interest_id`) REFERENCES `interests` (`id`);
And i have only about 30,000 records in the database but all queries which use this table is consuming alot of time, is it a design problem? note that type_id is linked to more than one table.
You should index your table properly, more information here: http://dev.mysql.com/doc/refman/5.7/en/mysql-indexes.html
It depends on your server aswell, you can "tune" your MySQL with this script: https://www.howtoforge.com/tuning-mysql-performance-with-mysqltuner
I'm using a barcode scanner to scan qr codes from tickets for an event, this will be posted to a database with the following structure:
CREATE TABLE IF NOT EXISTS `attendees` (
`id` int(11) NOT NULL AUTO_INCREMENT,
`attendeeguid` varchar(32) NOT NULL,
`scanned` int(1) NOT NULL DEFAULT '0',
`date` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP,
PRIMARY KEY (`id`),
UNIQUE KEY `attendeeguid` (`attendeeguid`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1 AUTO_INCREMENT=15 ;
Now this attendeeguid has to be unique, but if an event is 2 or more days, it should be possible to scan this attendeeguid again on another date. How can I do this?
I'm using MySQL and I'm using this in an Objective-C iOS App.
I have archived some old line items for invoices that are no longer current but still need to reference them. I think I need to create a VIEW but not really understanding it. Can someone help so I can run a query to pull the invoice and then the total of all the line items assigned (no matter what table the items are in)?
CREATE TABLE `Invoice` (
`Invoice_ID` INT(11) UNSIGNED NOT NULL AUTO_INCREMENT,
`Invoice_CreatedDateTime` DATETIME DEFAULT NULL,
`Invoice_Status` ENUM('Paid','Sent','Unsent','Hold') DEFAULT NULL,
`LastUpdatedAt` TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
PRIMARY KEY (`ID`),
KEY `LastUpdatedAt` (`LastUpdatedAt`)
) ENGINE=MYISAM DEFAULT CHARSET=latin1
CREATE TABLE `Invoice_LineItem` (
`LineItem_ID` INT(11) UNSIGNED NOT NULL AUTO_INCREMENT,
`LineItem_ChargeType` VARCHAR(64) NOT NULL DEFAULT '',
`LineItem_InvoiceID` INT(11) UNSIGNED DEFAULT NULL,
`LineItem_Amount` DECIMAL(11,4) DEFAULT NULL,
`LastUpdatedAt` TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
PRIMARY KEY (`LineItem_ID`),
KEY `LastUpdatedAt` (`LastUpdatedAt`),
KEY `LineItem_InvoiceID` (`LineItem_InvoiceID`)
) ENGINE=MYISAM AUTO_INCREMENT=1 DEFAULT CHARSET=latin1
CREATE TABLE `Invoice_LineItem_Archived` (
`LineItem_ID` INT(11) UNSIGNED NOT NULL AUTO_INCREMENT,
`LineItem_ChargeType` VARCHAR(64) NOT NULL DEFAULT '',
`LineItem_InvoiceID` INT(11) UNSIGNED DEFAULT NULL,
`LineItem_Amount` DECIMAL(11,4) DEFAULT NULL,
`LastUpdatedAt` TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
PRIMARY KEY (`LineItem_ID`),
KEY `LastUpdatedAt` (`LastUpdatedAt`),
KEY `LineItem_InvoiceID` (`LineItem_InvoiceID`)
) ENGINE=MYISAM AUTO_INCREMENT=1 DEFAULT CHARSET=latin1
Typically I would just run the following query to get the amount due on the invoices
SELECT
Invoice_ID,
Invoice_CreatedDateTime,
Invoice_Status,
(SELECT SUM(LineItem_Amount) AS totAmt FROM Invoice_LineItem WHERE LineItem_InvoiceID=Invoice_ID) AS Invoice_Total
FROM
Invoice
WHERE
Invoice_Status='Sent'
Also how can I select all the line items from both tables in one query?
SELECT
LineItem_ID,
LineItem_ChargeType,
LineItem_Amount
FROM
Invoice_LineItem
WHERE
LineItem_InvoiceID='1234'
You can use the MERGE Storage Engine to create a virtual table that's the union of two real tables:
CREATE TABLE Invoice_LineItem_All
(
`LineItem_ID` INT(11) UNSIGNED NOT NULL AUTO_INCREMENT,
`LineItem_ChargeType` VARCHAR(64) NOT NULL DEFAULT '',
`LineItem_InvoiceID` INT(11) UNSIGNED DEFAULT NULL,
`LineItem_Amount` DECIMAL(11,4) DEFAULT NULL,
`LastUpdatedAt` TIMESTAMP NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
KEY (`LineItem_ID`),
KEY `LastUpdatedAt` (`LastUpdatedAt`),
KEY `LineItem_InvoiceID` (`LineItem_InvoiceID`)
) ENGINE=MERGE UNION=(Invoice_LineItem_Archived, Invoice_LineItem);
You can use UNION :
SELECT a.* FROM a
UNION
SELECT b.* FROM b;
You just need to have the same number and type of column in your different queries.
As far as I remember, you can add test in sub-queries, but I'm not sure you can order on the global result.
http://dev.mysql.com/doc/refman/4.1/en/union.html