我正在尝试从9 GB以上大小的表中获取数据,并拥有数百万条记录。我正在用该数据填充DataTable。我从表中以块的形式获取记录,即通过Ajax和SQL Limit查询每页10条记录。
在上图中,您可以看到我们有server="localhost"
页,因此当我尝试访问最后一页时,查询将花费很多时间来加载数据。但是,当我尝试访问首页时,数据加载会更快。但是直接访问更高偏移量的页面需要花费很多时间。
223,740
PHP函数/控制器:
public static function getAllEvaluationsWithNameForDataTable($start){
$queryBuilder = new Builder();
return $queryBuilder
->from(array('e' => static::class))
->leftJoin('Cx\Framework\Models\Common\User\CxUser', 'e.cx_hc_user_id = u.id', 'u')
->columns('e.id, e.first_name, u.initials as assigned_coach, e.gender, e.email, e.phone, e.age, e.version, e.evaluation_status, e.ip_address, e.date_created, e.date_updated')
->orderBy('e.id asc')
->limit(10, $start)
->getQuery()
->execute()
->toArray();
}
Javascript:
public function getEvaluationsAction() {
// Enable Json response
$this->setJsonResponse();
// This action can be called only via ajax
$this->requireAjax();
// Forward to access denied if current user is not allowed to view evaluation details
if (!$this->CxAuth->currentUserIsAllowedTo('VIEW', CxEbEvaluation::getClassResourceName()))
return $this->forwardToAccessDeniedError();
if(isset($_GET['start'])){
$start = $this->request->get('start');
}else{
$start = 10;
}
$recordsTotal = count(CxEbEvaluation::getAllForDataTable(array('id')));
//Get Evaluations from DB
$evaluation_quizzes = CxEbEvaluation::getAllEvaluationsWithNameForDataTable(intval($start));
//for getting base URL
$url = new Url();
$data = array();
foreach ($evaluation_quizzes as $key => $quiz) {
$data[ $key ][ 'id' ] = $quiz[ 'id' ];
$data[ $key ][ 'first_name' ] = $quiz[ 'first_name' ];
if($quiz[ 'assigned_coach' ]){
$data[ $key ][ 'assigned_coach' ] = $quiz['assigned_coach'];
}else{
$data[ $key ][ 'assigned_coach' ] = "Not assigned";
}
$data[ $key ][ 'gender' ] = $quiz[ 'gender' ];
$data[ $key ][ 'email' ] = $quiz[ 'email' ];
$data[ $key ][ 'phone' ] = $quiz[ 'phone' ];
$data[ $key ][ 'age' ] = $quiz[ 'age' ];
$data[ $key ][ 'version' ] = $quiz[ 'version' ];
$data[ $key ][ 'quiz' ] = $url->get('/admin/get-evaluation-quiz-by-id');
$data[ $key ][ 'manage-notes-messages-and-calls' ] = $url->get('/admin/manage-notes-messages-and-calls');
$data[ $key ][ 'date_created' ] = date("m/d/Y H:i:s", $quiz[ 'date_created' ]);
$data[ $key ][ 'evaluation_status' ] = $quiz[ 'evaluation_status' ];
}
// Return data array
return array(
"recordsTotal" => $recordsTotal,
"recordsFiltered" => $recordsTotal ,
"data" => $data //How To Retrieve This Data
);
// Return data
}
我希望问题清楚。如果需要其他任何内容,请告诉我。
(来自评论)
cx.common.data.cxAdminDataTables.EbEvaluation = $CxRecordsTable.cxAdminDataTable({
ajaxUrl: '<?php echo $this->CxHelper->Route('eb-admin-get-evaluations')?>' + eqQuizIdQueryString,
serverSide: true,
processing: true,
recordsFiltered :true,
columns: [
cx.common.admin.tableEditColumn('id',{ delete: true }),
{ data: 'first_name' },
{ data: 'assigned_coach' },
{ data: 'gender' },
{ data: 'email' },
{ data: 'phone' },
{ data: 'age' },
cx.common.admin.tableLinkColumn('quiz', quizLinkOptions),
cx.common.admin.tableEditColumn('id', healthCoachLinkOptions),
cx.common.admin.tableLinkColumn('manage-notes-messages-and-calls', manageNotesMessagesAndCalls),
{ data: 'date_created' },
cx.common.admin.tableSwitchableColumn('evaluation_status', {
editable: true,
createdCell: function (td, cellData, rowData, row, col){
$(td).data('evaluation-status-id', rowData.id);
},
onText: 'Complete',
offText: 'In progress'
})
],
toolbarOptions:{
enabled: false
}, success: function (data) {
cx.common.data.cxAdminDataTables.EbEvaluation.cxAdminDataTable("reloadAjax");
}
});
}
else {
$row.removeClass('alert');
}
});
}
});
答案 0 :(得分:4)
由Masivuye Cokile链接的Why does MYSQL higher LIMIT offset slow the query down?问题和答案以及此处提供的https://explainextended.com/2009/10/23/mysql-order-by-limit-performance-late-row-lookups/链接包含了一个很好的总结,说明了为什么大偏移量查询很慢。基本上,对于LIMIT 150000, 10
,MySQL仍会扫描整个150000行,即使稍后将其丢弃也是如此。要加快速度,您可以:
count
查询计算出的大致页码。id
上创建索引,然后强制mysql执行仅索引搜索。 对于第二种方法,您必须从
重写查询SELECT ...
FROM table t
WHERE ...
ORDER by t.id ASC
LIMIT 150000, 10
到
SELECT ...
FROM (
SELECT id
FROM table
ORDER BY
id ASC
LIMIT 150000, 10
) o
JOIN table t
ON t.id = o.id
WHERE ...
ORDER BY t.id ASC
或者,由于您不仅限于单个查询,因此可以使用
来检索页面上第一项的ID。SELECT id
FROM table
ORDER BY id ASC
LIMIT 150000, 1
然后使用上述ID检索实际数据:
SELECT ...
FROM table
WHERE id >= $id
AND ...
ORDER BY id ASC
LIMIT 0, 10
答案 1 :(得分:2)
模式SELECT whatever FROM vast_table ORDER BY something LIMIT 10 large_number
是一个臭名昭著的性能反模式。为什么?因为它必须检查很多行才能返回几行。
如果您的id
值是主键(或任何索引列),则可以按
SELECT whatever FROM vast_table WHERE id BETWEEN large_value AND large_value+9 ORDER BY id;
或者您可以尝试
SELECT whatever FROM vast_table WHERE id >= large_value ORDER BY id LIMIT 10;
如果您的id
值中有空格,这将不能完美地分页。但是它的表现可以接受。
答案 2 :(得分:1)
问题与我表中的dates列数据类型有关。我在日期字段中使用int
数据类型,而当我将日期列的数据类型更改为datetime
时,搜索结果以秒为单位。
我在http://dbscience.blogspot.com/2008/08/can-timestamp-be-slower-than-datetime.html找到解决方案的来源