用php导入10W条+ 级别的csv大文件数据到mysql。导出10W+级别数据到csv文件

转自：http://blog.csdn.net/think2me/article/details/12999907

1. 说说csv 和 Excel

这两者都是我们平时导出或者导入数据一般用到的载体。两者有什么区别呢？csv 格式更兼容一点。那么共同点都是GBK格式的，非UTF8。所以我们上传文件的时候，老是出现乱码，就是编码问题没有转好导致。

2. 推荐的几种方法

1. 函数 fgetss($handel); 返回字符串。它就是strip_tags(fget($handel))的组合读取csv 文件的一行。去掉了其中的HTML，php 等标签。这种方法使用于小的csv 文件

[php] view plain copy





$handle=fopen("1.csv","r");

while(!feof($handle)){

$buffer=fgetss($handle,2048);

$row = mb_convert_encoding(trim($buffer), 'utf-8', 'gbk'); //很重要。转换成UTF8格式，不然容易产生乱码

$data=explode(",",$row); //转换成数组

$insertRows[] = $data;

} //这样所有的csv文件就生成一个二维数组$insertRows;

2. 函数 fgetcsv($handel,2048,','),返回数组，它就是explode(",",fget($handel))的组合。这种方法使用于小的csv 文件。而且不适合有汉字的csv 文件。

[php] view plain copy





$handle=fopen("1.csv","r");

while($data=fgetcsv($handle,1000,",")){

$insertRows[] = $data;

} //这样所有的csv文件就生成一个二维数组$insertRows;

3. 说到正题了。上面2种方法适合少量的csv 文件。如果一个文件有10W+ 以上的行数，恐怕就很悲剧了。

[php] view plain copy





$handle=fopen("1.csv","r");

//将文件一次性全部读出来

$excelData = array();

$content = trim(file_get_contents($fileName));

$excelData = explode("\n",$content);

或者直接用$excelData = file($file); file(） 函数直接将数据读出并放入数组当中。

我们先将所有的文件一次性读出来。放到一个数组$excelData 中。这个时候，我们就可以不用管这个csv 文件了，纯粹了php 操作数组了。所以。不会崩溃异常：

[php] view plain copy





$chunkData = array_chunk($excelData , 5000); // 将这个10W+ 的数组分割成5000一个的小数组。这样就一次批量插入5000条数据。mysql 是支持的。

$count = count($chunkData);

for ($i = 0; $i < $count; $i++) {

$insertRows = array();

foreach($chunkData[$i] as $value){

$string = mb_convert_encoding(trim(strip_tags($value)), 'utf-8', 'gbk');//转码

$v = explode(',', trim($string));

$row = array();

$row['cdate'] = empty($v[0]) ? date('Y-m-d') : date('Y-m-d',strtotime($v[0]));

$row['business'] = $v[1];

$row['project'] = $v[2];

$row['shopname'] = $v[3];

$row['shopid'] = $v[4];

$row['fanli'] = formatNumber($v[5]);

$row['fb'] = $v[6] * 100;

$row['jifen'] = $v[7];

$sqlString = '('."'".implode( "','", $row ) . "'".')'; //批量

$insertRows[] = $sqlString;

}

$result = $model->addDetail($insertRows); //批量将sql插入数据库。

}

插入数据库当然是批量插入了：

[php] view plain copy





public function addDetail($rows){

if(empty($rows)){

return false;

}

//数据量较大,采取批量插入

$data = implode(',', $rows);

$sql = "INSERT IGNORE INTO tb_account_detail(cdate,business,project,shopname,shopid,fanli,fb,jifen)

VALUES {$data}";

$result = $this->query($sql);

return true;

}

ok ! 亲测试。10W 数据。6个字段。插入需要10秒。

-----------------2013-12-18 日更新---------------------

3. 导出10W+的数据到csv

放弃之前写的一篇博客中用到的方法：http://blog.csdn.net/think2me/article/details/8596833 。原因是：当超过50W+ 以上的数据时，有可能浏览器崩溃，内存超。

这个方法是写文件的方式。然后再把文件弹出下载。

[php] view plain copy





public function dump2Excel() {

set_time_limit(0);

ini_set('memory_limit', '640M');

//获取列表

$name = $this->getActionName();

$model = D(GROUP_NAME . '.' . $name);

$map = $this->_search();

//文件名

if (isset($_GET['error']) && $_GET['error'] > 0) {

$filename = C('IMG_PATH').'account_data_error_' . $map['action_id'] . '_' . date('Y-m-d', mktime()) . '.csv';

}else{

$filename = C('IMG_PATH').'account_data_all_' . $map['action_id'] . '_' . date('Y-m-d', mktime()) . '.csv';

}

//用户信息,商家ID,联盟,商家订单号,商品分类,确认类别,下单时间,完成时间,

//实际支付金额,佣金,佣金补贴,返利,F币,论坛积分,备注,强制入库

// 'user_info', 'shopid', 'league', 'order_id', 'classify', 'confirm_type',

//'buydate', 'paydate', 'real_pay', 'commission', 'commission_plus',

// 'fanli', 'jifen', 'bbs', 'remarks', 'persist_execute', 'unique_sign','error_code'

$header[] = iconv("utf-8", "gb2312", "用户信息");

$header[] = iconv("utf-8", "gb2312", "商家ID");

$header[] = iconv("utf-8", "gb2312", "联盟");

$header[] = iconv("utf-8", "gb2312", "商家订单号");

$header[] = iconv("utf-8", "gb2312", "商品分类");

$header[] = iconv("utf-8", "gb2312", "确认类别");

$header[] = iconv("utf-8", "gb2312", "下单时间");

$header[] = iconv("utf-8", "gb2312", "完成时间");

$header[] = iconv("utf-8", "gb2312", "实际支付金额");

$header[] = iconv("utf-8", "gb2312", "佣金");

$header[] = iconv("utf-8", "gb2312", "佣金补贴");

$header[] = iconv("utf-8", "gb2312", "返利");

$header[] = iconv("utf-8", "gb2312", "F币");

$header[] = iconv("utf-8", "gb2312", "论坛积分");

$header[] = iconv("utf-8", "gb2312", "备注");

$header[] = iconv("utf-8", "gb2312", "强制入库");

$header[] = iconv("utf-8", "gb2312", "唯一标识");

$header[] = iconv("utf-8", "gb2312", "错误信息");

$headerFile = implode(',', $header);

//写入标题

@unlink($filename);

file_put_contents($filename, $headerFile."\n");

//获取所有error_code

$list = D('Fanli')->table('tb_account_action_data_error_code')->field('id,err_msg')->findAll();

$error_msg = array();

foreach ($list as $value) {

$error_msg[$value['id']] = $value['err_msg'];

}

//导入错误的数据

if (isset($_GET['error']) && $_GET['error'] > 0) {

$map['error_code'] = array('gt', 0);

}

if (!empty($map['action_id'])) {

$allCount = $model->where($map)->field('count(1) as count')->select();

$pageLimit = ceil($allCount[0]['count']/self::PAGE_COUNT);

$voList = array();

//打开文件

if (!$handle = fopen($filename, 'a')) {

echo "不能打开文件 $filename";

exit;

}

//分页获取

for($i=0;$i<$pageLimit;$i++){

$count = self::PAGE_COUNT;

$start = $count * $i;

$limit = "$start,$count";

$voList = $model->where($map)->limit($limit)->order('id desc')->findAll();

//写入文件

$excelString = array();

foreach ($voList as $v) {

$dumpExcel = array();

$dumpExcel[] = mb_convert_encoding($v['user_info'], 'GBK', 'UTF-8');

$dumpExcel[] = mb_convert_encoding($v['shopid'], 'GBK', 'UTF-8');

$dumpExcel[] = mb_convert_encoding($v['league'], 'GBK', 'UTF-8');

$dumpExcel[] = mb_convert_encoding($v['order_id'], 'GBK', 'UTF-8');

$dumpExcel[] = mb_convert_encoding($v['classify'], 'GBK', 'UTF-8');

$dumpExcel[] = mb_convert_encoding($v['confirm_type'], 'GBK', 'UTF-8');

$dumpExcel[] = "'".mb_convert_encoding($v['buydate'], 'GBK', 'UTF-8');

$dumpExcel[] = "'".mb_convert_encoding($v['paydate'], 'GBK', 'UTF-8');

$dumpExcel[] = mb_convert_encoding($v['real_pay'], 'GBK', 'UTF-8');

$dumpExcel[] = mb_convert_encoding($v['commission'], 'GBK', 'UTF-8');

$dumpExcel[] = mb_convert_encoding($v['commission_plus'], 'GBK', 'UTF-8');

$dumpExcel[] = mb_convert_encoding($v['fanli'], 'GBK', 'UTF-8');

$dumpExcel[] = mb_convert_encoding($v['jifen'], 'GBK', 'UTF-8');

$dumpExcel[] = mb_convert_encoding($v['bbs'], 'GBK', 'UTF-8');

$dumpExcel[] = mb_convert_encoding($v['remarks'], 'GBK', 'UTF-8');

$dumpExcel[] = intval($v['persist_execute']);

$dumpExcel[] = mb_convert_encoding($v['unique_sign'], 'GBK', 'UTF-8');

$dumpExcel[] = mb_convert_encoding($error_msg[$v['error_code']], 'GBK', 'UTF-8');

$excelString[] = implode(',',$dumpExcel);

}

//只能一行行些。不然容易漏

foreach($excelString as $content){

fwrite($handle, $content . "\n");

}

unset($excelString);

}

fclose($handle);

}

//导出下载

header("Content-type: application/octet-stream");

header('Content-Disposition: attachment; filename="' . basename($filename) . '"');

header("Content-Length: ". filesize($filename));

readfile($filename);

}